CORRECTIONS IN EARLY
QUR?AN MANUSCRIPTS
TWENTY EXAMPLES
DANIEL ALAN BRUBAKER
Copyright © 2019 by DANIEL A. BRUBAKER
Published by Think and Tell Press, Lovettsville
All rights reserved.
Front cover: MS.474.2003, courtesy of the Museum of Islamic Art (D. Brubaker
photograph); back cover: BnF arabe 331, Bibliothéque nationale de France
Photographs of manuscripts from the following institutions are used by
permission:
e Bibliothéque nationale de France, Paris
e Museum of Islamic Art, Doha
e The National Library of Russia, St. Petersburg
Modern Qur’an images are from the Mushaf Muscat, used by permission of
DecoType.
No part of this book may be reproduced in any form or by any electronic or
mechanical means, including information storage and retrieval systems,
without written permission from the author, except for the use of brief
quotations in a book review.
Corrections in Early Qur’an Manuscripts: Twenty Examples, Quran Manuscript
Change Studies (series) Vol. 1
ISBN13: 978-1-949123-05-0 (paperback, full color edition)
CONTENTS
Transliteration Key
Manuscripts Referenced
Preface
Acknowledgments
1. Introduction
2. The Corrections
3. Conclusions
Index of Quran verses referenced
Further Reading
Glossary
About the Author
103
105
109
113
TRANSLITERATION KEY
For Arabic transliteration, I use the IJMES system, shown
below. However, since my purpose is to emphasize the form of
script rather than its pronunciation, I don’t change the definite
article lam to the sound of the “sun” letters immediately
following it.
xu Transliteration Key
Sum
sh (3
SUS
dus
tL
zs
"e
ghé
fis
qa
kJ
lJ
ome
no
he
W 3
y¢?
-a (-at in construct state) 6
al- and ’1- (article) J!
alores
Transliteration Key xt
When I need to transliterate the bare rasm, I use Thomas
Milo’s system of capital letters to convey the ambiguity of
archigraphemes in manuscripts:
l 9
= %
G ‘a em K J 4s
-— J
&)
D > L J
R B 2 M e
S uw yf pin o J
C oe S= H é é
T L b& wy 3 2S
E ‘a ro [B]-y Ss
MANUSCRIPTS REFERENCED
Bibliothéque nationale de France, Paris
arabe 327 (examples 7 and 9)
arabe 328 (examples 2 and 12)
arabe 330 (example 8)
arabe 331 (example Io)
arabe 340 (example 13)
Cairo Al-Hussein Mosque
Cairo mushaf al-sharif (example 16 and “Another
phenomenon’ at the end of chapter 2)
Dar al-Makhtutat, San‘a’
O1-20.4 (example 3)
Museum of Islamic Art, Doha
MIA.2013.19.2 (example 15)
MIA.2014.491 (example 20)
MS.67.2007.1 (example 6)
xu Manuscripts Referenced
MS.474.2003 (example 5, and front cover)
National Library of Russia, St. Petersburg
Marcel 2 (example 4)
Marcel 5 (example 19)
Marcel 7 (example 18)
Marcel 11 (example 3 and 17)
Marcel 13 (example 3)
Marcel 21 (example 3)
Topkap1 Palace Library, Istanbul
Topkap1 mushaf al-sharif (examples 1, 11, and 14)
PREFACE
Over the several years since defending my doctoral disserta-
tion, “Intentional Changes in Qur’an Manuscripts,” I’ve
received many inquiries about when this work will be
published. The delay has been the result of several factors,
foremost among them a personal tendency toward perfec-
tionism that keeps me revising in what sometimes feels like an
endless loop. Another is the sheer, and growing, volume of the
material that I have accumulated.
The book in your hands is a small attempt in the short
term to give some satisfaction to those eager for this work by
providing a representative sample and introduction. It is
unusual, I think, for the popular-level publication to precede
the full academic one, and I have been cautioned by some
against doing this. I feel this situation is a special case — and
at any rate I am happy to release this material even as I finalize
the larger and more rigorous works.
In the following pages, I’ve written to be understood not
only by scholars but also by general readers, though without, I
xviii Preface
hope, compromise of academic integrity. It is not a big book,
but it does include some technical details, as it should. If you
are a non-specialist, there may be things that go over your
head. Please do not worry about it too much. Conversely, if
you are an academic, you may find yourself wishing for more
detail. If this is the case for you, I do hope you will at least
derive some benefit from what I have done here, but also wait
patiently as I finalize further publications.
~~’
This book does not engage in theological analysis of the
Qur’an’s content. Though theology is a thoroughly legitimate
subject of study and serious contemplation, I do not comment
in these pages upon the ultimate spiritual questions that the
Qur’an raises. This book is not the place for it. What I have
done is introduce an aspect of textual criticism of the Qur’an
that has occupied my interest and effort and that I find
fascinating.
Still, it may be appropriate to consider why the Qur’an as
an object is worthy of such attention, what difference it should
make to most people that the book exists, and why regular
folks might be curious about its early transmission history.
The first thing to note is that the Qur’an’s ideas and
theology have not only had an impact on world history for
more than a thousand years, but they continue to affect the
lives of billions of people today. Those affected include people
who believe Muhammad was a prophet and those who do not.
Let me address several potential categories of readers, one
or more of which may represent you, and discuss what profit
there may be in the subject of this book for each.
Preface xix
First, if you are one who believes Muhammad was a
prophet and you call yourself “Muslim:” The Qur’an is, quite
simply, your kitab, your book. Obviously, you know this. When
it comes to these manuscripts, they are among the earliest
surviving witnesses to the message Muhammad delivered. The
verses they contain are likely those to which Muhammad was
referring when he is reported to have said that Allah “showed
favor to the believers when He sent among them an apostle
from among themselves who recited to them His verses and
purified them and taught them the book and wisdom, though
before they were in obvious error.”! For this reason alone |
suggest it makes sense to have a curiosity about the early
manuscripts. What passages do they contain? What are their
physical and textual features? What variations exist among
them, and what can these mean? What are these corrections,
and why do they exist? And so forth. These manuscripts are
fascinating to me as a scholar, and they capture my imagina-
tion every time I handle them or think about where and when
they came into being. I can only imagine how special they
must be to a person who believes they contain records of
revelation.
Second, what might be the interest for those who do not
believe Muhammad was a true prophet and therefore do not
believe the Qur’an to be revelation from God? To this I would
answer that, even so, there are indeed some 1.6 billion people
in this world who do believe that Muhammad was a prophet
and who thus live their lives in some degree of conformity to
his message and instructions. Depending upon where you live,
you may have a lot of direct interaction with such people, or a
little, or perhaps none at all. But the world is changing
constantly and the influence of ideas upon events distant and
xXx Preface
local is all around us. It is probably not necessary for you (if
you fall into this category) to engage deeply with the literary or
devotional sources of the religion of Muhammad, but a
reasonable level of information — some of it on the cutting
edge of research — about the history and documentary trail of
the foundational text of this world faith, the Qur’4an, is worth
acquiring, if possible, and some of that is contained in the
pages that follow here. Also, in my view, the main contours of
the book you hold in your hands are not terribly difficult to
grasp, even for non-Arabists. Give it a go!
Third, if you are an academic (either believing or non-
believing), the purpose of your work is to uphold truth and to
pursue knowledge. Yes, it is true that everyone has a reason
and motivation that draws our interest to a particular subject
— you do as well as I. But to be a good scholar, one must
operate in methodologically sound ways, and in service of a
common master: Truth. When it comes to the history of
Muhammads life and actions, to that of those who identified
with him during that time and in the subsequent Arab
conquests, to the history of the revelation and its transmission,
and to the content of that revelation, the circumstances of its
delivery, and the exegetical (interpretive) keys that may have
been given by Muhammad and related through his compan-
ions and others, there is a lot to unpack and a lot more uncer-
tainty about some details than most uncritical observers may
understand. As scholars we intend to test claims and proposi-
tions, with the purpose of refining our understanding of what
actually happened or (as the case may be) what was actually
said or written. Today we are witnessing an exhilarating time,
academically, for the study of Qur’a4n manuscripts and their
early history. As an academic who is presumably interested in
Preface XX1
this field (since you are reading this), you may be very
encouraged.
Fourth, if you are a professional (either believing or non-
believing) in politics or government or media, you may be
faced from time to time, and perhaps increasingly, with claims
about Islam, both positive and negative. Depending upon your
personal relationships, predilections, political leanings, or any
number of other factors, you may be inclined to believe one or
the other. But in any given situation you might be wrong or
lack nuance. When you speak, people listen to you. When you
make a policy decision, it affects others. Good governance,
reporting, and managing of others rests ideally upon knowl-
edge and wisdom. Sometimes the critical or negative thing —
even though not politically correct — might be true. At other
times, the alarming generalization — even though satisfying
to believe — might not be true or may at least require some
qualification. The very center of Islam is Muhammad, and the
center of Muhammad’s identity for Muslims is his status as a
Messenger of Allah. A messenger has a message, and the
Qur’an is that of Muhammad; therefore, its history is not an
inconsequential matter.
This book, as I have already said, will not speak to the
larger issues of Islam. But it does focus on tangible historical
objects that, because of their particular characteristics as
described here, do challenge traditional assertions about the
transmission of the Qur’an in several ways. If you are a
curious and inquisitive person (and these are really good and
important qualities in any human being, all the more so for
those whose professions are related to knowledge, policy, and
opinion), this book might be of interest to you not only for
what it says about the matter at hand but also for what it says
XXL Preface
about what might be termed a pious enhancement of the
Qur’an’s textual history. Hagiography, the enhancement of a
history in order to elevate its subject, is not only carried out on
historical people; it can be directed toward objects as well and
— setting aside completely for the moment the entire matter
of the actual nature of the Qur’an as received by Muhammad
himself — the history of this recitation as a physical object
from the time of its writing down until the present appears to
contain some elements of hagiography, or cleaning up and
beautifying. To acknowledge this is not necessarily to suggest
bad or nefarious motives and certainly it is quite natural for
people to attribute the most complimentary attributes and
circumstances to people or objects that they hold in great
esteem. But if it is our intention to deal in reality, as is or ought
to be the case for historians and reporters and legislators alike,
we should be willing to test assumptions.
Sy”
The past dozen years have been an adventure for me. I became
interested in Qur’an manuscripts while working on my PhD
in the department of Religious Studies (now the department
of Religion) at Rice University in Houston. In fact it was at a
conference in Oxford around that time, perhaps 2007, that I
first heard Keith Small present a paper on textual criticism of
the Qur’an. I had further conversations with Keith on the
subject at that time, and he, in his characteristic kindness and
humility, made himself available to mentor me as I became
ever more interested over the next several years.
Keith invited me to deliver a paper as part of a panel at the
annual meeting of the Middle East Studies Association
Preface xxi
(MESA) in San Diego, organized by Emran El-Badawi and
chaired by David Powers. I had been examining photographs
of Qur’an manuscripts (for example from the UNESCO CD-
Rom of San‘a’ Qur’ans) more closely, and found a very inter-
esting page with two corrections and one glaring non-canon-
ical variant that had not been corrected. I discussed these
findings in my conference paper.
During a break between sessions at that conference, Keith
and I were talking in the sitting area of his hotel room in the
Grand Hyatt (it looked right out over the Bay and the day was
sunny and beautiful) and Keith showed me a couple of
photographs that really caught my attention: they were
pictures of Qur’a4n manuscript pages that had quite dramatic
and lengthy corrections on them. These were not just a few
letters or a single word, but were overwritten erasures greater
than a full line in length. I found the photographs fascinating
and surprising.
Already very interested, I decided to look more at the early
development of the written Qur’an in my doctoral work,
thinking at first that I wanted to write my dissertation on
something in the area of textual criticism related to these
manuscripts generally, including corrections as part of the
picture. As I began this project, I began to find more correc-
tions, and I took note of them. In 2011 I made a major research
trip to Europe and the Middle East to view manuscripts.
Although my original idea for my dissertation was to write on
early development of the Qur’an in written form, I began to
think seriously about writing only on corrections. I contacted
Keith once again and asked his opinion of this direction. He
said it would be a very good topic, so I told David Cook and
my department of the new subject I intended to pursue.
xxiv Preface
The rest is now history. I successfully defended my disser-
tation and was awarded my doctorate from Rice in April 2014,
and have been continuing my research in this area since.
Whereas I had documented some 800 physical corrections in
my dissertation, I have by now noted thousands, and there is
no end in sight.
What do these corrections indicate? You will see my own
brief remarks in the final chapter.
SS’
I am aware, of course, that my work deals with things that real
people believe, feel in their hearts, and consider to be matters
of cultural and personal honor, so it is appropriate that I speak
about this aspect fora moment.
The matter of corrections in Qur’an manuscripts obvi-
ously touches the question of whether what we have now is a
true and complete representation of what was delivered by
Muhammadin the first part of the 7th century AD.? This ques-
tion is quite different from the (also important) one of whether
Muhammad was a prophet — that is, whether these revela-
tions are from God. The book you have in your hands, and the
material contained within it, does not have anything to say
about whether Muhammad was a prophet. It does have to do
with questions about the Qur’an’s original form and about the
integrity of its transmission in the earliest stages after Muham-
mad’s death. I am not trying to hurt anyone's feelings by
studying these things or by talking about them. What I would
like to do, both as a human being and as a scholar, is test
assumptions and follow evidence where it leads. I propose this
Preface XXU
path is a good one for anyone to follow, and so J invite you to
come along this road as well.
As mentioned earlier, my more extensive works are forth-
coming; I do hope those interested will wait patiently a little
while longer and receive these warmly when they are soon,
God willing, published.
At the back of this book is a list of further reading.
Daniel Alan Brubaker
May, 2019
1. Guillaume, A., The Life of Muhammad: A translation of Ibn Ishaq’s Sirat Rastl
Allah, (Oxford: Oxford University Press, 1955. 398). See also Qur’an 3:164.
2, A note on dating conventions: I reference dates throughout this book in the
Gregorian (solar) calendar, that is, “AD” and “BC.” Many readers will be
aware that there is also an Islamic lunar calendar that dates from the year
in which Muhammad and his community emigrated from Mecca to
Medina, 622 AD, an event called the hijrah. Its dates are given in (lunar)
years as “AH,” for Anno Hegirae, “year of the Hijrah.” In many scholarly
books dealing with subjects related to Islam and its history, dates are given
in both AD (or “CE”) and AH. For simplicity and ease of reference to the
calendar familiar to most readers, I have chosen not to do this in the
English edition. If readers wish to find corresponding hijri dates in a given
instance, there exist today many free phone apps and online calculators
that make doing so very easy.
ACKNOWLEDGMENTS
Many people have made my Qur’an manuscript work to date
possible and pleasant, but there are some in particular whom I
would be remiss not to mention by name.
First, I am grateful to my Ph.D. advisor David Cook and the
late Keith Small, both academic mentors to me, and both also
dear friends. I am also grateful to have had the guidance and
friendship of the late Andrew Rippin, and am particularly
thankful for the honor of his presence on my doctoral
committee.
Many owners, curators, caretakers, and staff of institutions
housing the manuscripts have opened doors for me, and I am
grateful to all, including the following: Olga Vasilyeva and the
manuscript staff at the National Library of Russia; Sue
Kaoukji, her team at the Dar Museum (Kuwait); Dr. Mounia
Chekhab Abudaya, Marc Pelletreau, and the entire staff at the
Museum of Islamic Art in Doha; Marie Geneviéve Guesdon at
the Bibliothéque nationale de France; Alasdair Watson and
the staff of the Special Collections Reading Room at Oxford
University; Catherine Ansorge and Yasmin Faghihi at the
Cambridge University Library; Elaine Wright at the Chester
Beatty Library (Dublin); Colin Baker for making available to
me both the actual manuscript BL2165 as well as a personal
copy of his facsimile edition of the same in September of 2013
at the British Library; Samar A] Gailani of the Beit Al Qur’an
as well as its curator, Ashraf A] Ansari; Dr. Halit Eren in
Istanbul for his hospitality in 2011 and to IRCICA and Dr.
Tayyar Altikulag (who I have yet to meet) for his wonderful
work in preparing the Turkish facsimile editions of important
Qur’an masahif; ISAM; the Muslim Board of Uzbekistan for
their assistance on my 2016 visit to Tashkent; and Amalia
Zhukovskaya and Alla Sizova at the Institute of Oriental
Manuscripts in St. Petersburg for their help during my several
visits there.
Colleagues who have offered help and/or hospitality in this
work include Efim Rezvan, Gerd-R and Elisabeth Puin (who
hosted me as a houseguest and have invited me again), Alba
Fedeli, and Francois Déroche.
I give a friendly nod also to colleagues at both the Islamic
Manuscript Association (particularly Davidson MacLaren)
and the International Qur’anic Studies Association (particu-
larly Emran E]-Badawi).
Latha and I owe a personal debt to Joshua Lingel for his
steady encouragement over many years. He is a dear friend.
In 2012, having become aware of his unique expertise and
long work in the area of historically and scientifically sound
Arabic typography and encoding, I began pursuing Thomas
Milo to develop much-needed solutions specific to my own
needs dealing with Qur’an manuscripts, in particular related to
the un-disambiguated rasm.' We had conversations around this
topic over the following two years, and the conversation contin-
ues. I am heartened today to see fruit coming in large part
through the clear outside-the-box thinking and aesthetic sense
possessed by him and Mirjam, who have become friends to us.
Iam so grateful. They are kind, smart, and talented people.
I am grateful to my friend and colleague Roy Michael
McCoy III (Ph.D. as of January 2019) who in this post-doctoral
stage has assisted immensely with, among other things,
transfer of my research from notes and photographs to the
database I designed for the purpose of organizing and
containing this material; it has been a privilege to work along-
side him and my other colleague and friend, Andy Bannister -
whose talents and energies seem to know no end.
The following people read and offered helpful peer-review
comments on this book, in more than one instance saving me
from a great deal of embarrassment. I owe each of them
sincere thanks: Marijn van Putten, Gerd-R. Puin, Asma Hilali,
and Mark Durie. Further, I thank Leah Garber for her thor-
ough proofreading for grammar, punctuation, and style.
Finally, I thank my dear (also talented, accomplished, kind
and beautiful) wife, Latha, for being my partner in life’s work. I
also thank my parents, Alan and Susan Brubaker, as well as
hers, Annamma and the late Jefreys K. Samuel, for bringing us
into the world, for raising us with much love and sacrifice, and
in particular for helping during the long hours and days of my
doctoral work and beyond.
Having said all this, the work that follows is my own and I
take responsibility for shortcomings that persist. I hope it
proves for all readers an enjoyable and informative introduc-
tion to a fascinating topic.
1. There is a glossary at the back of this book defining specialized terms,
including this one. The rasm, as I use the term, is the basic shape of the
Arabic consonantal text, without its dots or short vowels.
1
INTRODUCTION
arly Qur’a4n manuscripts contain many physical
changes or corrections.' By now I have taken note of
thousands of such changes via careful examination
of these manuscripts, mostly in person. This book is meant to
serve as an overview, providing examples to illustrate the
general nature of these manuscript corrections. In subsequent
works I will make a more extensive presentation of these
corrections and their descriptions.
ABOUT EARLY QUR’?AN MANUSCRIPTS
A wealth of Qur’an manuscript fragments have survived from
the first several centuries of its life. Many of the important
early manuscripts are now available to scholars like myself as
a result of such political circumstances as Napoleon’s
campaign in Egypt and Syria which was accompanied by the
deployment of scholars such as Jean-Joseph Marcel,” or as the
result of the intrepid efforts of people like Jean-Louis Asselin
2 DANIEL ALAN BRUBAKER
de Cherville,> Agnes Smith Lewis and Margaret Dunlop
Gibson,* Chester Beatty,” Edward Henry Palmer,® and others,
who acquired and preserved these objects. There have indeed
been many different people involved in the collection and
safekeeping of these manuscripts and these are only a few of
the important names behind the objects in western academic
libraries today. There are more manuscripts safely (others,
sadly, have been imperiled by wars and other political instabil-
ities) kept in institutions around the world, and these all have
stories and individuals behind them. It has taken me time and
effort over a number of years to learn the locations of many
such manuscripts, and I have had the privilege now of visiting
a large number, including those probably produced in the 7th
and 8th centuries AD, in libraries and museums around the
world.
With some very important exceptions — most notably the
Qur’ans that tradition tells us the third Caliph, ‘Uthman,
burned and that would therefore be forever lost — we seem to
have a good number of early Qur’an manuscripts from a fairly
early date. Why are there so many Qur’an manuscripts in
existence from the first and second centuries after Muham-
mad’s life? In addition to the relative recency of this revelation’
when compared to (for example) the biblical writings, there
are two additional main reasons.
First, by the 7th century AD (Muhammad, the sources tell
us, lived from 570 to 632), parchment was a commonly used
material to receive writing, particularly for books.’ Parchment
is animal skin, and unlike papyrus which will typically disinte-
grate over a period of 100-200 years, documents written on
parchment can endure for thousands of years. They do not
always last this long as other factors, such as the acidity of the
Corrections in Early Qur’an Manuscripts 3
ink used for writing, the quality and thickness of the parch-
ment itself, and the humidity and other environmental factors
the document has endured, contribute to their lifespan. But in
general, the relative stability of parchment has resulted in a
very large number of Qur’an manuscripts from these impor-
tant early centuries having survived for us to examine.
A second major reason for the survival of so many Qur’an
manuscripts from the early period is the fact that, beginning
in the mid-7th century AD (that is, the first century after
Muhammad) these were produced in political environments
that viewed the book in a positive light. The ruling authorities
in these regions were not hostile to the Qur’an as was the situ-
ation for the New Testament during the first two centuries of
Christianity. It was not dangerous to own a copy of the Qur’an
in the Arab empires that by the latter part of the 7th century
stretched across a huge swath of territory from the Iberian
Peninsula and the Maghreb in the West to Azerbaijan in the
East, nor were these manuscripts usually destroyed if discov-
ered. In fact, to own such an object was a sign of status, wealth,
and piety. These documents were finely produced and at great
cost. Professional scribes were employed and good materials
used, the best that could be afforded by the person who
commissioned a copy. As time progressed, the production of
Qur’ans became an art in itself with precise rules of geometry,
ruling, and letter form. IJluminations were applied in various
colors including gold leaf. These were treasured possessions
that held a place of prominence and were displayed openly in
mosques, palaces, and private residences.
4 DANIEL ALAN BRUBAKER
HOW MANUSCRIPTS ARE DATED
The first question people tend to ask when looking at one of
these manuscripts is, “How old is it?” Obviously, we want to
know when a manuscript was produced, because its date
allows us to (a) better understand what the object can tell us
about its time, and (b) apply what we may already know about
that period as a lens to help us understand what we see going
on in the manuscript itself. So, the date is very important.
It would be nice if every manuscript came with a label
stating when it was produced. In fact, it became customary in
later Qur’an manuscripts to include a colophon with such
information as the name of the scribe and the date of produc-
tion. Unfortunately, we do not have such neat and clean indi-
cations for manuscripts of the first several centuries.
These manuscripts, then, are dated by considering what
information we do have, and this typically includes things
such as paleography, or the study of the development of script
styles. We have a good sense of when particular styles of script
and certain developments in the ways of writing Arabic were
in use, and so this detail of a manuscript is very important.
The script style classifications that are standard today were
described by Francois Déroche in the 1980s. In general, the
script styles listed in rough chronology of their first appearance
are: “hijazi” or “ma?il” (these two terms are used interchange-
ably), O, A & B.Ia (similar time of origin), C, B.I, D, E, F, and
“New Style.”? These styles do overlap; one was on the rise as
another was still in use or in decline, for example, and even
this statement only takes into account the chronological
dimension; doubtless regional and economic factors play into
the picture as well. Most of these styles have subcategories. It
Corrections in Early Qur’an Manuscripts 5
is not an exact science — for example, it is common for a
manuscript to match Prof. Déroche’ description of a style
along most but not all of its defining features — but this
should surprise no one when it is remembered that the scribes
were human (scribes’ personal styles are most readily
observed in the earlier scripts) and that there was a progres-
sion over time and geography.
The earliest Qur’4n manuscripts, particularly those in the
“hijazi” or “ma’il” styles, were written without diacritic marks
or with only occasional diacritic marks to disambiguate the
archigraphemes. This is not to say, however, that the only way
to disambiguate the archigraphemes was via diacritics. In fact,
there came to be a system of writing the Arabic rasm that
allowed precise disambiguation without those extraneous
marks, and Thomas Milo has termed this system “script
grammar.”!°
A second helpful indication can be the features of the page
or book beyond the writing itself. The study of these features
is called codicology.''! Codicology asks questions like these:
What is the writing material? Is the page vertical or horizontal
in format? What are the dimensions of the page or book? How
many lines of writing are on each page? Do all pages have the
same number of lines, or does the number vary? How are the
verses and chapters divided, and what sort of marks are used
to do so? What inks were used? Is the page illuminated with
illustrations or other graphic elements, including extra mark-
ings to represent the short vowels? If so, what colors were used
and what forms or styles or particular types of elements are
present? Are the lines of the pages ruled? Are the margins
ruled? Do margins exist or does the writing extend to the edge
of the page? How is (or was) the book sewn together? What
6 DANIEL ALAN BRUBAKER
sort of binding was used? These features and more can
provide details that give additional clues as to the age of a
manuscript.
A third method for dating is probably the most well-
known: radiocarbon dating. This method can be applied to
anything organic. Everything that was once living, that is, all
plant or animal material, is organic. Parchment qualifies, and
can therefore be tested with this method. The reason radio-
carbon dating works is that a radioactive isotope of carbon is
present in all living things and begins to slowly decay at a
predictable rate when the living thing dies. Subjecting parch-
ment to this testing yields a series of date ranges based upon
the probable time that the source (in this case most likely the
goat or sheep) was alive.
Obviously a radiocarbon date range cannot tell when a
parchment was written, but we generally assume that a parch-
ment did not sit for decades before receiving its first writing.
Radiocarbon dating is not a foolproof way of determining
dates of manuscripts. Some manuscripts of known date of
origin (e.g. with a colophon or some other overt indication of
the time of writing) have been radiocarbon dated a century or
more off from the apparent actual date of production. There-
fore, all these methods must be taken with a grain of salt, and
in most cases the best thing is to take all the various clues
(paleography, codicology, and radiocarbon dating if available),
and weigh them alongside one another.
PROVENANCE
In dealing with any ancient artifact, we want to consider all
available information about it. At first view, we may be
Corrections in Early Qur’an Manuscripts 7
tempted to imagine that we can discern everything important
about a manuscript merely from its physical details — what it
says, what it does not say, how it was written, how it was orna-
mented, the material upon which it was written and the ink(s)
that were used, how it was bound, whether the page was ruled,
and so forth.
However, an object’s context can also be very important to
historians. There is always a context in which an item was
produced, and there were contexts through which these
objects have passed all the way until the time of their
discovery (or re-discovery) and even after their discovery.
Unfortunately, we cannot time-travel back to the moment and
place of production, so it is very helpful to know at least where
an item was found, by whom, and the chain of custody since
that time. Where a manuscript was found may give further
indication about where it was produced, during what time
period, and by whom, as well as to how it was used post-
production. Most of the manuscripts discussed in this book
were not recently found on archeological digs, but were rather
discovered in mosques or libraries or private family collections
passed down through generations and at some point (for
example) making their way into the stall of a flea market or
antiquities dealer and then sold to a discerning buyer. Still,
chain of custody is important for several reasons, including
the authentication of an object in a world where the value of
such items leads sometimes to forgeries. We certainly don’t
want to base our historical research about the past upon
objects that are not authentic.
I don't go deeply into provenance in this book, but will say
that much of the modern history of manuscripts shown is
documented, and that institutions will often express caution
8 DANIEL ALAN BRUBAKER
about objects whose provenance has not been confirmed or is
doubtful. Furthermore, once an image such as a photograph is
made of an object, that image also becomes an object. Who
took the picture, where, and when? Citing the photographer,
whether he or she is a museum staff member or a researcher
like myself who was permitted access, should always be done
when this information is available. This is both to give credit
where due and also to describe the context of the object with
due diligence.
It is important, finally, to understand that doubtful prove-
nance does not mean that an object isn’t authentic or that it
should not be taken seriously. Nor does the existence of a solid
chain of custody always mean that an ancient object must be
authentic, though it does strengthen the case. Attention to
provenance is merely one of the best practices in archeology
that helps us do quality work and avoid making unnecessary
errors.
CONSONANTAL VARIANTS
Leaving aside for a moment the matter of corrections to the
page, there is variation in the consonantal text (in Arabic this
is called the rasm) within early Qur’an manuscripts. The tradi-
tional way of accounting for this variation is to claim that it
was a flexibility approved by Muhammad himself and repre-
sented in variant readings, called qira’at. In fact, the readings
are different from the rasm and in most cases the one is not
affected in the least by the other. Indeed, the history of the
codification or (if you will) canonization of these readings is
more complex and, according to the recent work of Shady
1
Nasser,!? owed less to a historical root validating each of the
Corrections in Early Qur’an Manuscripts 9
particular readings than it did to pragmatic or practical and
political concerns. In short, Nasser argues, the readings were
chosen to give geographic representation during Ibn Mujahid’s
time (late 9th and early 1oth century AD) in the various urban
centers from which he chose them, and not necessarily on the
basis of strongest multiple attestation as is commonly
supposed.
A further matter of difficulty for the readings is that the
consonantal texts of some of the important monumental early
codices, such as the Topkap1, Istanbul, and Cairo masahif, do
not reflect a single reading, but rather what might appear to be
a combination of the different readings.'* This fact leads the
preparer of their facsimile editions, Dr. Tayyar Altikulag, to
describe these codices in terms of rough percentages when it
comes to their adherence to the various readings. Such a
circumstance is not necessarily irreconcilable with the exis-
tence of approved readings, but it does indicate a more
complex picture that requires further inquiry and explanation.
That being said, many of the thousands of corrections |
have documented appear to have nothing to do with the read-
ings attested in the secondary literatures. So, corrections must
represent in at least some cases another phenomenon, such as
perhaps a greater degree of perceived flexibility of the Qur’an
text in its early centuries (the time of first production of these
manuscripts) than is documented in the gird ’at literature.
GENERAL OBSERVATIONS ABOUT CORRECTIONS
You will see details of corrections in the next chapter, but will
not get a full sense of the relative prevalence of different types
of correction or their other features, so here is a high-level
10 DANIEL ALAN BRUBAKER
view. In a correction, something is added (insertion), removed
(erasure), replaced (erasure overwritten, taping overwritten, or
overwriting without erasure), or (perhaps) hidden. Corrections
can be classified in other ways but these terms summarize the
mechanics of the matter. I discuss the last category briefly at
the end of the next chapter.
Most of the time, I have found that corrections in a Qur’an
manuscript result in conformity of that manuscript at the
point of the correction with the rasm of the now-standard 1924
Cairo edition. This pattern is important and shows a general
movement over time toward conformity, though not imme-
diate complete conformity. There are interesting questions
raised when a manuscript is corrected in one place but
remains deviant (the word “deviant” supposes a standard and I
use it here merely as a practical matter) when compared to the
1924 Cairo edition in other places. We will discuss this scenario
more later.
Sometimes a correction takes a manuscript away from
conformity with the now-standard rasm. The first thing to
consider when that is observed is whether the correction has
followed a regional variant, and for this possibility there is a
secondary literature to consult. Very rarely, a correction actu-
ally takes the manuscript out of conformity with any docu-
mented variant or reading, so such instances are interesting
when they are found.
Because each correction is different in nature and signifi-
cance, it would be a mistake to draw conclusions from raw
numbers, but for general information, here is a rough break-
down of the relative number of instances so far.!°
e Erasure overwritten — about 30%
Corrections in Early Qur’an Manuscripts II
e Insertion — about 24%
e Overwriting without erasure — about 18%
e Simple erasure — about 10%
e Covering overwritten — about 2%
e Covering — about 16%
More important factors to consider than the mechanics of
a change include the apparent reason(s) for it, its timing rela-
tive to that of the first production of the manuscript, its extent,
and what has been altered. Relevant questions about these
matters, and more, will be discussed at the end of this intro-
duction.
WHERE ARE THESE MANUSCRIPTS TODAY?
Because of factors such as the climate of the region of their
production and the material upon which they were generally
written (parchment), a very large number of early Qur’an
manuscripts have survived the centuries and exist in private
and public collections. My own work over the past dozen years
has been an exciting process of discovery of (among many
other things) where they are. I now have a very long list and in
my travels have noted collections both large and small.
Undoubtedly, there are many of which I have yet to become
aware — including those in private collections.
But in terms of a general understanding, which is the point
of this book, these manuscripts exist in various university and
national libraries, such as Cambridge, Oxford, the University
of Birmingham, the John Rylands Library in Manchester, and
Berlin, as well as in museums around the world such as the
Museum of Islamic Art in Doha, the Tareq Rajab Museum and
12 DANIEL ALAN BRUBAKER
the Dar Museum in Kuwait, the British Library in London, the
Chester Beatty Library in Dublin, and the Beit al-Qur’an in
Manama, and the Biruni Institute or Oriental Studies in
Tashkent, to name a few. Thanks is also owed to private donors
and collectors, such as Nasser D. Khalili, who have gathered
and safeguarded these objects and who make them available
to scholars for study.
HISTORY OF THE QUR’AN (TRADITIONAL)
What is commonly accepted about the early history of the
Qur’an has reached us mainly through secondary literatures
that were written down beginning in the closing decades of
the 8th century (that is, about 150-160 years after the death of
Muhammad). These literatures, though further removed in
time from the events they describe than we might hope, are
not without merit: but different scholars and historians have
approached them in different ways. I will discuss this matter
more later; the first thing is to relate a general outline of the
traditional account, that is, what most Muslims and most
casual observers accept as “what happened.” Here it is:
Muhammad was born in 570 AD, in Mecca. His father died
before the time of his birth and his mother died when
Muhammad was yet very young. Henceforward, Muhammad
was raised by his grandfather and then by his uncle. As a
young adult, Muhammad entered the employ of a Meccan
businesswoman named Khadija, who was significantly his
senior. When he was 25, she proposed marriage and he
accepted.
At age 40, Muhammad was spending time alone in a cave
in the hills outside of Mecca where he sometimes went for
Corrections in Early Qur’an Manuscripts 3
quiet. Suddenly, he was encountered by an imposing being
that seemed to cover the sky. It grabbed him tightly and gave
the command “igra’!” (“recite”), to which he answered, “What
shall I recite?” Three times this happened, each time the being
grabbing him even tighter. After the third time, tradition tells
us, the first revelations — part of what is now the Qur’an —
began coming from his mouth.
Muhammad returned home in a sweat, not sure what had
just happened. It was his wife, Khadija, who informed him
that this had been the angel Gabriel and that Muhammad was
a Messenger of God.
This first encounter was in 610 AD. Over the next 22 years
(23 or so by the lunar calendar), Muhammad would continue
to receive revelations from time to time. Sometimes they were
long and sometimes short. Sometimes they were at close inter-
vals and at other times long periods elapsed between revela-
tions. When Muhammad would get a revelation, he ordinarily
would begin reciting it publicly, for example in his prayers.
Others among those who had become believers would listen,
memorize, and recite as well, thus recording and transmitting
the revelations orally. There are also hadith traditions that say
Muhammad would have his personal secretary, Zayd b.
Thabit, write down the revelations whenever he received
them.
By the time of Muhammad’s death in 632 AD, the revela-
tions had been, we are told, written down on various discrete
objects like palm stalks, stones, and bones of animals. These
were gathered together, compiled and organized around this
time, and written asa book (Arabic mushaf).
Over the next couple of decades, later sources state, there
came to be disagreements over some parts of the Qur’an’s
14 DANIEL ALAN BRUBAKER
content that were significant enough to require the resolution
of the matter via production of authoritative copies and the
destruction of those deemed variant. This process, we are told,
was undertaken by the third Caliph, ‘Uthman, who died in 656
AD. He commissioned the production of several authoritative
copies and had these sent out to the main centers of the now
large Arab empire that he oversaw.
‘Uthman’s suppression of variants is not the end of the
story, of course, even for the time period of the manuscripts
which are considered in this book. These manuscripts go up
through the 9th or possibly 1oth centuries. We don’t need to
cover all that history here, but I should mention some major
developments. By the closing years of the 7th century, the
Arabs had conquered territory all the way from Azerbaijan in
the East to the center of the Iberian Peninsula (via North
Africa) in the West. By “conquered,” we mean they had gained
political control over the regions, not that they had settled or
saturated the countryside or territories. The religion of
Muhammad filtered out more gradually and organically into
these areas over the subsequent decades and centuries.
There were rivalries and dynastic changes that occurred
both regionally and across time. We don’t need to cover all of
these here either; some highlights include the start of the
Umayyad dynasty of caliphs with the death of ‘Ali, the fourth
caliph who was also the cousin and son-in-law of Muhammad,
in 661 AD. The Umayyads held authority over most of the
Arab kingdom until the Abassid revolution in 750 A.D., and
the Abassids, though shifting capitals (Baghdad, Kufa,
Samarra, etc.) held sway until the mid-13th century.
Corrections in Early Qur’an Manuscripts 15
DIFFICULT ISSUES
Some aspects of the Qur’an, and some aspects of the historical
records of the larger context of its transmission, including the
details of the people and events in the first century of Islam,
are a puzzle to historians. For the most part the Qur’an’s
language is not complicated. However, it contains words and
phrasings that seem to have been inscrutable even to believing
exegetes going back to the early centuries of its history. Among
these are words at whose meaning even the early commenta-
tors have had to guess.
For a book claiming to be revelation from God to contain
mysteries, of course, would not itself be surprising. Some
people, however, have raised the question of how such a
circumstance can be reconciled to the Qur’an’s internal claim
to have been “revealed in a clear (mubin) Arabic tongue.”
(Q16:103) !°
Devin Stewart, in considering words that break the rhyme
structure of a passage, has entertained the possibility that the
rasm at some time came to be mis-pointed in places by a later
generation who did not have the benefit of an unbroken and
complete oral tradition.'’ Such a theory, if true, would alter
traditional assumptions about the Qur’an’s transmission
history. In any event, extensive re-visiting of the received
pointing of the Qur’an text, that is, major revisionism, is prob-
ably not warranted. Still, I think it is entirely appropriate to
consider the text in the way Devin has, and the rhyme words
would just be markers that highlight a larger phenomenon. If
it has happened with words that should rhyme, it would be
unreasonable to think that it did not happen elsewhere as well
— and the logical next thing to consider would be words that
16 DANIEL ALAN BRUBAKER
today pose difficulty for exegetes or that seem to be out of
place. Could the rasm be read in a way that makes sense but is
outside the received tradition of reading? The question has
been considered. !®
Following are a few more examples that highlight inter-
esting questions and issues that critical scholars have been
trying to address in recent years:
1) The topography and other features of Mecca does not
seem to match descriptions in the Qur’an. The Qur’an itself is
not rich in narrative detail, but this does not mean descrip-
tions are lacking entirely. When one looks closely, there are
solid things that may be observed. For example, the late
Patricia Crone noticed the agricultural details in Q36 mention
grain, date palms, and grapes, as well as gushing springs, with
some echoes of these agricultural references appearing also in
Q56. These descriptions are tied to the local pagans whom
Muhammad was being instructed by Allah to warn. She notes
many other agricultural references, most of which seem quite
disconnected from the reality of Mecca.'?
2) The archaeology of Mecca does not seem to support the
traditional assertions that the place where Muhammad grew
up and received the revelations was a location that had seen
the rise and fall of many previous civilizations.
3) The linguistic features of the Qur’an, in the opinion of
some linguists”? but not that of others,”! raise questions about
its place of origin. These are not, perhaps, questions irrecon-
cilable at this point with the broad outlines of the traditional
narrative, but neither are they insignificant.
4) The qibla, or direction of prayer, is designated by the
orientation of the wall of the mosque containing its mihrab,
the niche in the wall designating this direction. One recent
Corrections in Early Qur’an Manuscripts 17
researcher, Dan Gibson, has noted that the surviving founda-
tions of all the earliest mosques until about 706 AD do not
point toward Mecca at all, but rather considerably further
north,’* and such does indeed seem to be the case. After 706,
he finds giblas began pointing in a direction further south
from the original direction but still north of Mecca, and the
first gibla he found pointing toward Mecca dates to around 727
AD. In fact, some process of development of the direction of
prayer is attested in the literatures of the time, with some indi-
cations that the direction was at first merely toward the east,2°
though these sources diverge from others that indicate
Muhammad designating the gibla first toward Jerusalem and
then toward Mecca at a specific moment during his lifetime.“
Time will tell where scholarship lands on this matter as more
attention is directed toward reconciling the archaeology with
the contemporary historical literatures and other sources.
The apparent difference between what the Biography of
Muhammad (written by Ibn Ishaq and revised by Ibn Hisham)
says on this matter and what is seen in the mosque founda-
tions highlights a larger and quite well-known issue that will
be mentioned again later: the reliability of the existing
secondary literatures, such as histories, hadith collections,
biographical reports, and so forth. There is an extensive litera-
ture in Arabic from the 8th and 9th centuries relating history
of the previous century, but the documents contain internal
disagreements, sometimes without a clear clue for deter-
mining which side (if any) of a conflicting account is true. It is
not unusual to find equally “reliable” reports that are contra-
dictory.?°
5) The manuscripts support some aspects of the tradi-
tional narrative, such as the approximate time period during
18 DANIEL ALAN BRUBAKER
which quranic materials came to be written (for example, we
have portions of Qur’an manuscripts that appear to date from
the mid-7th century), and they often confirm the existence of
many of the various readings that are attested in the secondary
literatures of the following century, but other features present
a puzzle and need some sorting out.
First, many manuscripts do not follow a single reading but
rather appear — to a person who is operating from the point
of view of the documented canonical readings — to move
between readings without a discernible pattern. This is not a
problem but rather emphasizes the question, “What was the
place of the readings at the time of production of these manu-
scripts?”
Second, there exist entire pages of parchment that have
been washed or otherwise cleared of quranic material and
then rewritten. These sheets, called palimpsests, are the most
extensive corrections that have reached us. The wonderful
thing about these documents is that in many cases what was
first written on these pages can be discerned, either with the
naked eye or through the use of technology that picks up the
earlier text. | have not emphasized these in my research since
others like Alba Fedeli, Elisabeth Puin, Asma Hilali, Eléonore
Cellard, Behnam Sadeghi, and Mohsen Goudarzi have been
working with them, but I will reference these as appropriate in
later works.
Third, given the monumental nature of what tradition
reports the third caliph, ‘Uthman, to have done with the stan-
dardization of the text — the suppression of variants via
burning or other means of destruction, and the production of
authoritative copies that were then to serve as exemplars and
standards against which subsequent copies could be measured
Corrections in Early Qur’an Manuscripts 19
— it is odd that no copy existing today has been reliably iden-
tified as one of these actual authoritative copies, and that the
ones about which such a claim is made seem to have been
produced long after ‘Uthman’s time. Certainly there is
evidence in the lower text of the aforementioned palimpsests
that there were earlier forms of the text, but this does not solve
the problem of the apparent lack of any of ‘Uthman’s copies
existing today. These documents would have been extremely
important objects, so we would expect they would have been
preserved.
Fourth, the existence of manuscripts that were finely
produced yet sometimes corrected after a long passage of time
is interesting and presents a challenge to the notion that there
was a Strict uniformity and widespread agreement about every
detail, every word and letter, such as one would expect to find
if there were widespread agreement upon a standard from a
very early date, such as the time of ‘Uthman’s caliphate. I will
discuss some of my thoughts about this matter in
“Conclusions.”
WHY WERE CHANGES MADE?
Not all manuscript corrections are equal; each has a context
and a situation involving time, place, writing materials, envi-
ronment, exemplar, scholar, scribe, and so forth.
The most obvious cause that any one of us can easily
imagine if we put ourselves in the place of working as a scribe
is making a simple mistake when copying or writing, realizing
the mistake, and then correcting it soon after.
A simple mistake-and-correction scenario fits what we see
in some Qur’an manuscript corrections where the ink, nib,
20 DANIEL ALAN BRUBAKER
and writing style appear to match that of the rest of the page.
However, it does not fit all of them. In many cases there are
clearly other factors at play. Here are some of the questions |
ask that help me think carefully about what is going on in a
given situation:
e Is there a discernible reason that could have
caused a simple mistake? One of the common
reasons for mistakes in manuscript transcription
from an exemplar, for example, is the repeated
occurrence of a word or sequence of words in close
proximity to each other. A scribe may finish
copying the first instance of the word or word
sequence, go to dip the nib into the ink, and
accidentally begin writing again after the second
occurrence of the word or word sequence. This
could be noticed later and corrected. Such a
scenario or others like it is not uncommon in
manuscript transmission.
e Was there a long passage of time between first
writing and the moment(s) of correction? This
question can be pursued further by asking some of the
following:
1. Does it look as though the writing materials (ink
and nib width, for example) used in the correction
were similar to those used for the first production?
2. Is the writing style different from that of the main
page? Is it a later script style, one that became
popular in another time period? Is it of a different
dimension (e.g., taller or shorter), or is it ofa
Corrections in Early Quran Manuscripts
different nib angle, or is it written by a person ofa
different writing or skill level?
3. Is there a difference of orthography (that is, the
spelling or writing conventions that we know
developed over time) between the page as first
written and the part that has been corrected? Is the
correction itself possibly dealing with suchan
issue?
Here are some further questions to consider:
e Does the page show signs of having been corrected
more than once, at different times?
e What did the correction do? Can we see or guess
what was first written?
e What was the result of the correction? Is the
corrected rasm in harmony with the rasm of the
standard text today? If it is not, and if the nature of
the variant can be attributed to different
orthographic norms, does its orthography align
with other manuscripts from the time period? If it
does not, or if the difference cannot be attributed to
variant orthography, does it align with any of the
variant readings acknowledged in the gira ‘at
literature?
e Ifthe page has been corrected, what does the rest of
the page look like? Are there other parts of the page
that remain out of conformity with the now-
standard rasm, and if so, what could this mean
about the time that this document was corrected or
about the person who corrected it?
22 DANIEL ALAN BRUBAKER
Obviously, there are other questions one could ask, but
you can, I hope, begin to see the way that we try to unpack
these materials and make sense of them. You will see these
questions in action as we turn now to the main substance of
this book, and you can also ask them yourself as you look at
each example.
1. I use the term “correction” for convenience, but I ask readers to please
notice that the word itself carries a value judgment that I don’t necessarily
intend in each case. Was that which was first written necessarily less “cor-
rect”? Is what now stands always and necessarily more “correct”? Most of
the time, the changes we find in Qur’an manuscripts result in something
that looks more or less like the rasm of the standard 1924 Cairo text, but
there are exceptions. So, please bear in mind that when I use the term “cor-
rection,” I intend only to mean a physical change to the script at some
point.
2. Déroche, Frangois, Qur’ans of the Umayyads: A first overview, (Leiden: Brill,
2014), 17.
3. Ibid.
4. Ansorge, Catherine, “Cambridge University Library Islamic Manuscript
Collection. Origins and Content,” Journal of Islamic Manuscripts 7 (2016): 139-
40; Soskice, Janet, The Sisters of Sinai: How Two Lady Adventurers Discovered
the Hidden Gospels, (New York: Alfred A. Knopf, 2009). The latter tells the
fascinating story of how these two Scottish women traveled to North Africa
and procured important biblical and quranic manuscripts that are today
preserved in places like the Cambridge University Library.
5. A. Chester Beatty was a successful American businessman who used his
wealth for many charitable endeavors, among which was the acquisition of
manuscripts and other historic objects. Among the treasures he collected
were some of the oldest papyrus fragments of the New Testament, and
many Qur’an fragments and complete manuscripts, some quite early. Most
of his collection today is housed in the Chester Beatty Library, located in
the Dublin Castle.
6. Ansorge, Catherine, “Cambridge University Library Islamic Manuscript
Collection. Origins and Content,” Journal of Islamic Manuscripts 7 (2016): 135.
7. I do not intend by this comparison to suggest that the Qur’4n belongs in
the same category, qua revelation, as the Hebrew Bible and New Testament.
I merely raise it in order to point out similarities and differences in the
Il.
13.
14.
orrections in Ear ur?an Manuscripts 2
C t Earl -an M t
context and circumstances that may have factored into the outworking of
the various transmission histories.
. Papyrus was indeed also in use extensively during this time period.
Because of its lower cost, it was the writing material of choice for many
administrative and transactional documents, as well as regular correspon-
dence. Indeed, there are also examples of Qur’an manuscripts written on
papyrus. The few that I have seen (there are a couple in Oxford’s Bodleian
Library, for example) are small fragments. To my understanding, there are
a fair number of Qur’an manuscripts on papyrus, but | have not yet had
the opportunity to understand their number or quality. The point here is
that parchment use was widespread, and it is, at least in part, due to this
fact that we today have so many well-preserved pages of early Qur’ans.
. Déroche, Frangois, The Abbasid Tradition: Qur’ans of the 8th to the roth
centuries AD, (London: Nour Foundation, 1992); Déroche, Francois, Qur’ans
of the Umayyads: A first overview, (Leiden: Brill, 2014).
. Milo, Thomas, “Towards Arabic historical script grammar through
contrastive analysis of Qur’an manuscripts,” in Writings and Writing: Inves-
tigations in Islamic Text and Script in honour of Januarius Justus Witkam. Kerr,
Robert and Thomas Milo, eds. (Cambridge: Archetype, 2013), 249-92.
Déroche, Francois, Islamic Codicology: an introduction to the study of manu-
scripts in Arabic script, (London: Al-Furgan Islamic Heritage Foundation,
2006).
. Dutton, Yasin, “An Umayyad Fragment of the Qur’an and its Dating,” in
Journal of Quranic Studies 9, no. 2 (2007): 57-87.
Nasser, Shady, The Transmission of the Variant Readings of the Qur’an: The
Problem of Tawatur andthe Emergence of Shawadhdh, (Leiden: Brill, 2012).
What appears to be a combination of other attested readings, of course,
may be simply its own reading.
. These rough figures are mostly from my own work but also take into
account some corrections in several manuscripts found by my friend and
former research assistant, Dr. Roy Michael McCoy III. There is a great deal
of additional material in my own notes and photographs not yet included
in these numbers, and no doubt others will add to this body of research in
the future, but at this point I] do not expect a major shift in the relative
proportions.
. Ibn Kathir, Isma ‘il, Tafsir al-qur an al- ‘azim, (Beirut: Dar el-Marefah, 2003),
894-5. The commentaries (of which Ibn Kathir is but one that is somewhat
of a culmination taking into account the earlier historical and exegetical
sources) place this verse in the context of accusations against Muhammad
that he had been taught the Qur’an by someone else, in particular a
foreign servant who spoke only a little Arabic. This verse, then, is seen by
24
18.
19.
20.
21.
22.
23.
24.
DANIEL ALAN BRUBAKER
the commentators as an answer in which is implicit a rhetorical question:
“How could a foreigner be the source of verses composed in pure Arabic?”
. Stewart, Devin J., “Divine Epithets and the Dibacchius: Clausulae and
Qur’anic Rhythm,” Journal of Qur’anic Studies 15.2 (2013): 22-64. For further
discussion of rhyme as an organizing feature, see Sinai, Nicolai, The
Qur?an: A historical critical introduction, (Edinburgh: Edinburgh University
Press, 2017), 16-20.
Luxenberg, Christoph, The Syro-Aramaic reading of the Koran: A contribution
to the decoding of the language of the Koran, (Berlin: Verlag Hans Schiler,
2007); Bellamy, James A., “Some Proposed Emendations to the Text of the
Koran,” Journal of the American Oriental Society 113, no, 4 (1993); Bellamy,
James A., “More Proposed Emendations to the Text of the Koran,” Journal of
the American Oriental Society 116, no. 2 (1996).
Crone, Patricia, “How did the quranic pagans make a living?” Bulletin of
SOAS 68, no. 3 (2005): 387-399-
Durie, Mark, The Qur’an and its biblical reflexes: Investigations into the genesis
of a religion, (Lanham: Lexington Books, 2018), 16-17, 42-43 (note 22).
Nicolai Sinai (ibid., 42-43); van Putten, Marijn, “Hamzah in the Quranic
Consonantal Text,” in Orientalia 87 no. 1 (2018): 93-120.
Gibson, Dan, Qur’anic Geography: A survey and evaluation of the geographical
references in the Qur’an with suggested solutions for various problems and issues
(Altona: Independent Scholars Press, 2011).
Sharon, Moshe, “Qibla Musharriga and early Muslim prayer in churches,”
in The Muslim World Vol. LXXXI, Nos. 3-4 (1991).
“And when the qibla was changed from Syria to the Ka‘ba — it was
changed in Rajab at the beginning of the seventeenth month after the
apostle’s arrival in Medina — Rifa‘a b. Qays; Qardam b. ‘Amr; Ka‘b b. al-
Ashraf; Rafi‘ b. Aba Rafi‘; al-Hajjaj b. ‘Amr, an ally of Ka‘b’s; al-Rabi b. al-
Rabi‘ b. Aba’l-Hugayq; and Kinana b. al-Rabi‘ b. Abu’l-Hugayq came to
the apostle asking why he had turned his back on the gibla he used to face
when he alleged that he followed the religion of Abraham. If he would
return to the gibla in Jerusalem they would follow him and declare him to
be true. Their sole intention was to seduce him from his religion, so God
sent down concerning them: “The foolish people will say: What made them
turn their back on the gibla that they formerly observed? Say, To God
belongs the east and the west. He guides whom He will to the straight path.
Thus we have made you a central community that you may be witnesses
against men and that the apostle may be a witness against you. And we
appointed the qibla which thou didst formerly observe only that we might
know who will follow the apostle from him who turns upon his heels,’ i.e.
to test and find them out. ‘Truly it was a hard test except for those whom
God guided,’ i.e. a temptation, i.e. those whom Allah established. ‘It was
25.
Corrections in Early Qur’an Manuscripts 2
5
not Allah’s purpose to make your faith vain,’ i.e. your faith in the first qibla,
your believing your prophet, and your following him to the later gibla and
your obeying your prophet therein, i.e. so that he may give you the reward
of both of them. ‘God is kind and compassionate to men.” Guillaume, A.,
The Life of Muhammad: A translation of Ibn Ishaq’s Sirat Rasil Allah, (Oxford:
Oxford University Press, 1955), 258-9.
One example is what the historian al-Tabari relates about Muhammad’s
answer to the question about whether it was Isaac or Ishmael that
Abraham took up the mountain to sacrifice. Half the accounts say that
Muhammad answered, “Isaac,” and half say that he answered “Ishmael.”
Brinner, William M., tr., The History of al-Tabari, volume II: Prophets and
Patriarchs (Albany: State University of New York Press, 1987), 82-97.
2
THE CORRECTIONS
’ve chosen the following examples for this introduction
to the range of the phenomenon. | could easily have
chosen twenty others, and in subsequent installments
may do so. Included below are corrections of various types
(erasures, erasures overwritten, overwriting without erasure,
and insertion) as well as of different script styles representing
different early time periods (7th, 8th, and 9th centuries A.D.).
Since I understand that many readers of this book do not
speak or read Arabic, I’ve made an effort to explain each
change clearly in a way that will not be inscrutable to nonspe-
cialists. Translation and graphic elements should serve this
purpose while also including sufficient technical detail to
satisfy those who want it. There will still be difficult elements
for non-Arabic speakers, but I trust that the main point will be
understood from the photographs and the accompanying
descriptions.
28 DANIEL ALAN BRUBAKER
Example 1: Post-production insertion of a word in a
monumental 8th century Qur’4n
FIGURE |: Topkapi mushaf al-sharif, fol. 122v. (Source: Altikulag, Tayyar,
Ed. Al-Mushaf al-Sharif attributed to ‘Uthman bin ‘Affan (The copy at
the Topkap1 Palace Museum). Istanbul: IRCICA, 2007.)
This insertion is in the Topkapi: codex commonly known
as the Topkap: mushaf al-sharif. This codex, of 408 folios on
vellum, is distinguished as one of the oldest complete (two
folios are missing and others appear to have been replaced at
an early date) copies of the Qur’an. It was sent to Sultan
Mahmud II in 1811 as a gift by M. Ali Pasha, then Governor of
Egypt, and has been kept at the Topkap1 Palace Museum since
its arrival there in 1811.!
The Topkapi codex has been attributed by tradition to the
third Caliph, ‘Uthman, a Companion of Muhammad who
died in 656 AD, about 24 years after Muhammad himself died.
As is often the case with popular opinion, the attribution is not
correct; this codex probably dates to a century later, that is, the
mid-8th century AD. It is a delicate matter to challenge the
Corrections in Early Qur’an Manuscripts 29
attribution to ‘Uthman, so the statement of Mr. ihsanoglu, the
founding Director General of IRCICA and Secretary General
of the Organisation of the Islamic Conference, is admirable
and carries weight:
Judging from its illumination, the Topkap:1 Museum Mushaf
dates neither from the period when the Mushafs of the
Caliph ‘Uthman were written nor from the time when
copies based on these Mushafs were written. Since Mushafs
of the early period took those attributed to the Caliph
‘Uthman as a model, they do not have elements of
illumination. [...] This Mushaf [...] does not constitute a
sample of the early period of Mushaf writing due to a
number of characteristics [...It] most probably belongs to the
Umayyad period.”
The Topkapi codex is an important and _ beautifully-
produced relatively early and nearly complete monumental
Qur’an manuscript. I hope to discuss it further in subsequent
works.
I have noted 25 instances of correction over the Topkapi
mushaf al-sharif’s 408 folios. This, as well as examples 11 and 14,
are representative. The photograph above shows an insertion
of the word 54 huwa, “it [is],” of Qg9:72. In the 1924 Qur’an, the
affected phrase of this verse reads wa-ridwadnun mina Ilahi
akbaru dhdalika huwa °!-fawzu ’I-‘azimu “and Allah's good plea-
sure is greater, that is the great triumph.”
The words dhdlika huwa together mean “that is,” but
dhalika alone, which was part of the page as first written,
carries the same basic meaning. In other words, this particular
correction resulted in a manuscript that it is now in
30 DANIEL ALAN BRUBAKER
conformity with the now-standard rasm, but had no notable
semantic effect.
This is clearly a post-production correction. It has been
made with a different hand, nib, and style. It is my opinion
that there was the passage of some length of time between
production and correction.
% 2 filiaoes peal ‘ 7h) Nee pies a0 A
U x8 Bs SNe Ong SHAN 520 2 Ah)
.
Joo a |
9.9
FIGURE 2: IIlustration of location of correction Example 1 compared with the
Azhar-approved mushafmuscat.com Quran, which is based upon the 1952
Cairo edition. [The 1952 edition corrected some errors in the 1924 Cairo
edition. The mushafmuscat is basically the corrected 1924 edition with
Omani-style punctuation. Also, the mushafmuscat follows the Medina format
of 604 pages with a verse number at the end of each page for the entire
Quran, in contrast to the Cairo edition’s free flowing text over 827 pages. The
rasm is the same between these editions except for the position of some of the
amphibious characters. On the 1924 and 1952 editions, see Puin, Gerd-R,
“Quellen, Orthographie und Transkription moderner Drucke des Quran,” in
Vom Koran Zum Islam, Grof, Markus and Karl-Heinz Ohlig, Eds. 606-641.
Subsequent figures will refer to this as “the 1924 Cairo text” for simplicity.]
Corrections in Early Qur’an Manuscripts 31
Example 2: Post-production erasure overwritten in a 1st/7th
century Qur’an
Jal
thn Aa,
FIGURE 3: BnF arabe 328, fol. 58v.
This example is from BnF arabe 328b, part of the Codex
Parisino-Petropolitanus, which comprises BnF arabe 328a and
328b, as well as other folios that exist today in the National
Library of Russia in St. Petersburg, the Vatican Library, and the
Nasser D. Khalili Collection of Islamic Art in London.?
Francois Déroche has worked with this codex for many
years and has described it in wonderful detail. It dates, in his
opinion, to the third quarter of the 7th century, specifically to
between 671 and 695 AD.’ Dr. Altikula¢ similarly places it
(referring to arabe 328a) in the 7th century and, like Déroche,
believes it not to have been one of the copies produced by
‘Uthman, but rather a copy of one of them or a copy of a copy.
Although Déroche is cautious about placing it geographically,°
Altikulag sees evidence of origin in Damascus and suggests
32 DANIEL ALAN BRUBAKER
that was either copied from the codex that ‘Uthman sent
there, or from a copy based upon it.° The Codex Parisino-
Petropolitanus has many interesting distinctive features that
are beyond the scope of this book.
The photograph above shows a place where the page has
been erased and overwritten. Erasures were typically made by
scraping off the ink with a pumice stone; this process leaves
scratches on the parchment. It was often done in a very precise
way following the general shape of the letters that were erased.
Erasure marks are clear at this spot and J have looked at this
page on two different occasions. The change has been made by
a different hand and with a different nib and different ink than
the rest of the writing on the page. Among other things, the
lam (the upright extension at the right side) is less confident
and more vertical in contrast with the general rightward slant
of the rest of the page, including the apparent lam that was
erased.
This correction occurs at Q42:21, and is the second of three
instances of a4! lahum in this verse as it stands in the 1924 Cairo
edition. What was first written here appears to have been lam-
he, that is, the compound Arabic word lahu, “to him.” It has
been replaced by lam-he-mim, that is, lahum, “to them (m.).” As
such, the way this page was first written would have had a
meaning, “Or do they have associates who enacted for him as
a religion that for which Allah did not give leave?” instead of
the now-standard text which says “Or do they have associates
who enacted for them,” etc. The third person singular is used
in the previous verse, and as it was first written on this page,
verse 21 could have been read with the “for him” referring back
to the hypothetical individual mentioned in verse 20, who
wishes the tillage of the Hereafter.
Corrections in Early Qur’an Manuscripts 33
The way the page is written after this change corresponds
at this point with the consonantal text of the 1924 Qur’an. This
correction is not the only one on this page of the manuscript;
there are at least two others, including an erasure 3 lines
earlier.
Ao ees ape p
vA erasure
f ‘ x
I i. sli St tet Y
zo 7 , ples Capt eit ae ee alt
| yieineie erat
Ie] Ga goicdlal god Laslol Tee
v4 Op alias iy Sg jk ej Bi Y
FIGURE 4: Example 2 compared with the 1924 Cairo text
34 DANIEL ALAN BRUBAKER
Example 3: Multiple post-production insertions of “allah”
in several 1st/7th and 2nd/8th century Qur’ans
FIGURE 5: Nine insertions of allah in various manuscripts (Source of San ‘a’
image (bottom right): UNESCO CD of San ‘a’ Qur ’ans)
THE ABOVE FIGURE is not a single but a composite image
showing nine different manuscript insertions of the word
allah’ (“God”) at places where the word was omitted at the
time of the manuscripts’ initial production. I have so far cata-
logued about a dozen such instances in Qur’an manuscripts
produced in the 7th and 8th centuries, most of these in the
Fustat Umayyad Codex, and it has fascinated me to discover
that, of all things for a scribe to “forget,” allah would be among
them. I don't believe, actually, that allah was truly forgotten in
all these cases; in almost every instance shown above, Allah is
the implied subject but is not grammatically necessary. This
Corrections in Early Qur’an Manuscripts 35
recurrence of similar corrections in different places seems to
me evidence, perhaps, of a certain degree of early flexibility in
the manuscripts and probably also reflects the oral nature of
the transmission (since manuscripts are not produced in a
vacuum) that was at some later point in time drawn toward
uniformity.
Here is a description of each one of the above, from top left
to bottom right:
THE “FUSTAT UMAYYAD CODEX”
Many full manuscripts have been broken up into sections
and remain separated today in different libraries and muse-
ums. Such is the case with one codex that Francois Déroche
has named the Fustat Umayyad Codex. He believes it could
possibly be either the codex sent by al-Hajjaj to the ‘Amr
mosque, or that made by ‘Abd al-‘Aziz b. Marwan in response
to this action.’ Whatever the case, this codex apparently
remained in the ‘Amr mosque for a thousand years until the
early 1800s, when the various portions were acquired by Jean-
Joseph Marcel and making their way to Europe.? The
manuscript fragments now exist under four call numbers:
three in Russia (Marcel 11, Marcel 13, and Marcel 15) and one in
Paris (BnF arabe 330c).!° This manuscript is written in the O I
script style'! and was probably produced in the first part of the
8th century AD.
The Fustat Umayyad Codex has many interesting features.
Prof. Déroche has described it in in detail, and I have seen and
made my own close observations of all its folios in both Paris
and St. Petersburg. The interesting feature I am highlighting in
this example of change (which actually consists of seven
36 DANIEL ALAN BRUBAKER
changes in this codex and two in other ones) is the apparent
late standardization of a number of instances of allah. Here is a
description of each one in turn, with the manuscript and folio
listed as well as the particular instance of allah that was
omitted and then inserted shown in bold:
1. NLR Marcel 01, 7v. Q33:18, gad ya’lamu Ilahu ?I-mu ‘awwigqin
minkum, “Allah surely knows those from among you who
hinder others...” This is an erasure overwritten, but it is almost
certainly the allah that was missing earlier; if this was the case,
the ya’lamu was erased and both words were then written in.
As such, this manuscript prior to the change would have read,
“He surely knows those from among you who hinder others...”
2. NLR Marcel 11, 8r. Q33:24, li-yajziya Ilahu °I-sadiqin bi-sidqi-
hum, “In order that Allah might reward the truthful for their
truthfulness... .” Prior to the insertion, this manuscript read,
“In order that he might reward the truthful for their truth-
fulness.”
3. NLR Marcel 11, ov. Q33:73, wa-yatuba llahu ‘ala ’Il-mu’minina
wa-?l-mu’minat, “and that Allah might pardon the believing
men and believing women.” Prior to the insertion, this
manuscript read, “And that he might pardon the believing
men and believing women.”
4. NLR Marcel 11, 12v. Q4r:21, gala *antagana Ilahu lladhi antaqa
Corrections in Early Qur’an Manuscripts 37
kulla shay ’in, “they will say, ‘Allah who gave everyone speech
gave us speech.” Prior to the insertion, this manuscript read,
“they will say, ‘He who gave everyone speech gave us speech.”
5. NLR Marcel 13, 20v. Q22:40, yudhkaru fiha smu Ilahi kathiran,
“wherein the name of Allah is mentioned frequently.” Prior to
the insertion, this manuscript read, “Wherein the name is
mentioned frequently.”
6. NLR Marcel 13, 23r. Q24:51, du ‘U ila llahi wa-rasulihi, “they
are called unto Allah and his Messenger.” Prior to the inser-
tion, this page looks like a nonviable reading because of the
presence of the wdaw (“and”). It is therefore not clear what
might have been going on in this sentence.
7. NLR Marcel 13, 26r. Q35:11, inna dhdlika ‘ala lahi yasirun,
“that is indeed an easy matter for Allah.” Prior to this insertion
it is unclear how or whether this manuscript would have been
read sensibly at this point.
OTHER MANUSCRIPTS
8. NLR Marcel 21, 4v, line 11. Q9:93, wa-taba ‘a Ilahu ‘ala qulubi-
him, “and Allah has placed a seal upon their hearts.” Prior to
this insertion, the manuscript read, “and he has placed a seal
upon their hearts.”
38 DANIEL ALAN BRUBAKER
Marcel 21 is a horizontal fragment on parchment of 12
folios in 3 quires. It is a composite fragment, in that the third
quire, folios 9-12 (two bifolios), are clearly originally from a
different codex than the first two quires. As this insertion
comes from folio 4, I will only describe that portion of this
manuscript. Its pages measure about 17.9 cm tall by 29.5 cm
wide (about 7” x 11.6”), with the text block measuring 13 cm. tall
by 23 cm. wide. The script style is Déroche’s A.I. This first part
of Marcel 21 was likely produced in the early 8th century; the
third quire may date to the 7th century. Altogether, I have
noted about three dozen corrections in Marcel 21.
9g. UNESCO CD of San‘a’ Qur’ans, shelf number o1-20.4.
Q9:78, wa-’anna Ilaha ‘allamu 7I-ghuyub, “and that Allah
knows fully the things that are unseen.” Prior to this insertion,
this manuscript read, “and that he knows fully the things that
are unseen.
The final allah insertion shown in this example, from the
bottom right of Figure 5, is found on a page from the San‘a’
manuscripts. I do not have the dimensions of the page and
have seen it only in a photograph, not in person as I have all
the others in Example 3. The page is horizontal in format, with
22 lines on the page. It has virtually no margin, and in this
regard is very similar to the earliest vertical hijazi manuscripts,
which also tend to make full use of the page right out to the
edges. This is probably a late 7th or early 8th century
manuscript.
Corrections in Early Qur’an Manuscripts 39
ae ad
shill; <
Si Ossy alse ieee
asi sl; bt aye ah SHI ples
poof SS sioglly ale gs gil
ataiasiec tesa ome
fic (Ry Senate atl NN izes ibe
FIGURE 6: Illustration of the allah insertion at Q33:18 in the Fustat-
Umayyad codex as compared with the 1924 Cairo text
40 DANIEL ALAN BRUBAKER
ij
pp eleall, etalys Na ceil
ve
ep Ondine sts
LS) gael
iis @ a :
bse +, 3s M5 C9 bg Rearaey a Zs Sel
FIGURE 7: Illustration of the allah insertion at Q33:24 in the Fustat-
Umayyad codex as compared with the 1924 Cairo text
(
aan ee
NTE Oe BE Hil HH Ay gall [A
aes ie a hnctel
(
Oleh eatozosie,
FIGURE 8: Illustration of the allah insertion at Q33:73 in the Fustat-
Umayyad codex as compared with the 1924 Cairo text
Corrections in Early Qur’an Manuscripts 4I
es bps
WANs:
Syn Bee 1s JA ib 353)
ies wheal &, 28k cl RE $55 @
a ieerat Levees
FIGURE 9: Qgz:21 in the 1924 Cairo text, showing location of the allah
insertion in manuscript #4 above
Pe -
eee nails eo
Vlg BAS maar ae ©
FIGURE ro: Qz22:40 in the 1924 Cairo text, showing location of the allah
insertion in manuscript #5 above
42 DANIEL ALAN BRUBAKER
Exo
e ssl Sis ss ae: aha
pA i Alidpealne ally 56 Qa itl] 2
ola sites > S Salt |eyotecb ety gh E
FIGURE 11: Q24:51 in the 1924 Cairo text, showing location of the allah
insertion in manuscript #6 above
*
“2
FIGURE 12: Q35:11 in the 1924 Cairo text, showing location of the allah
insertion in manuscript #7 above
Corrections in Early Qur’an Manuscripts 43
4 Hi-allle ons ls Ay
So er oe
FIGURE 13: Q9:93 in the 1924 Cairo text, showing location of the allah
insertion in manuscript #8 above
| tsa pioats [f
i i cageelaa pine, 1G
d ieee dial ligt 4
4 TE Pere r
> one. sige waite Y
FIGURE 14: Illustration of allah insertion at Q9:78 in manuscript #9 above,
showing the location in the 1924 Cairo text
44 DANIEL ALAN BRUBAKER
THE ABOVE NINE corrections represent about three-quarters of
the simple allah insertions ] have noted so far. In addition to
these, there are many corrections beyond simple insertions
that involve allah. On its face, this should not be terribly
surprising, since allah is one of the most common words in the
Qur’an. Still, the specific nature of the corrections above
makes them worthy of attention.
Example 4: An erasure
ee 2] | ee
QOS ,
FIGURE 15: An erasure leaving a gap in Marcel 2, on the last line of the page.
This correction is found in the manuscript Marcel 2, in the
National Library of Russia, on folio 30v. This is a large square
Qur’an, with pages measuring about 41 cm. (~16 in.) square. Its
text block measures 33 cm. tall by 31 cm. wide (~13 in x 12 in.).
Its format is similar to that of the Cairo mushaf al-sharif.
Marcel 2 has 42 folios with 20-21 lines of script per page. It
contains verse dividers in the form of vertical stacks of diag-
onal nib marks, as well as multi-verse dividers in the form of
red medallions circled with brown ink, preceded by stacked
nib marks as mentioned already. Occasionally it has multi-
verse dividers in the form of a red medallion with four spikes
at diagonals and petals extending right and left, up and down.
These pages are written in the script style C.la, and this is
Corrections in Early Qur’an Manuscripts 45
probably an early 8th century manuscript. I’ve noted 26
corrections in this manuscript fragment.
The correction in this case is a simple erasure; nothing has
been written to replace what was erased. A gap is left by the
erasure at the end of the line, the final line on the page. The
erasure follows the word dae “qibatu, “the fate,” of Q30:9. The
word that comes next in the 1924 Cairo edition, ¢»!! alladhina,
“(of) those,” is the first word written on the following page of
this manuscript. So, the rasm now aligns at this point with the
1924 Cairo text.
This verse follows a narrative that chastises disbelievers for
not recognizing the signs and the fate of those who disobeyed
God in the past: “Have they not travelled in the land to see
what was the fate of those who preceded them? They were
stauncher than them in strength, and they plowed the earth
and built it up better than they themselves built it up, and
their messengers came to them with the clear proofs. Allah
would not wrong them but they wronged themselves.”
What was erased cannot at this time be discerned, but the
length and continuity of the erasure indicates a likely single
word of 4-6 letters, all linked. Grammatically, assuming that
the rest of the verse was read at the time of this manuscript as
it is today, there are possibilities that could fit in this space.
The first would be an expression of proportion such as kullu
min, “all of,” or kathiran min, “most of,” to render “what was the
fate of all those who preceded them,” or “what was the fate of
most of those who preceded them,” respectively. Another
possibility would be a noun (for example, al-yahid, “the Jews,”
or al-nds, “the people”) with the resulting translation, “what
was the fate of the Jews who went before them,” or “what was
the fate of the people who went before them.” To be clear, I
46 DANIEL ALAN BRUBAKER
have no indication that the erased writing said any of these
things; I mention them to illustrate that there are grammati-
cally viable possibilities.
a4 (Shs si b5 lil BBY giles
FAG Betod SO. lial
4 ts var
first word on
next pees
Lenya evil, ful
Ri aes fever his ate Nlee ee ees
SODA Be. Sasa ablail
FIGURE 16: Q3o:9 in the 1924 Cairo text, illustrating the Marcel 2 correction
—"
There is one other correction on this page in Marcel 2, an
insertion in the left margin. Like the one above, it has also
brought the page toward conformity at that point with the
rasm of the 1924 Cairo edition.
Corrections in Early Qur’an Manuscripts 47
Example 5: A Qur’an page, possibly 2nd/8th-3rd/9th
century, containing several post-production corrections
FIGURE 17: MS.474.2003, fol. 9v. (Brubaker photo, by permission of the
Museum of Islamic Art)
This page is in the Museum of Islamic Art in Doha, Qatar.
Its script style is Déroche’s A.I, and it is probably of 8th
century production. The manuscript fragment (MS.474.2003)
has about 30 physical changes over its 12 folios, and it is variant
when compared with the 1924 Cairo edition.
The photograph above (Figure 17) shows part of folio 9v, a
page that contains at least five instances of correction. Before
discussing these, here is a general description of the script on
this page.
This folio begins in the middle of Qé6:91. Even as it now
stands, it has a variant rasm; for example:
e wa-la “nor” of Q6:91 is written instead as waw,
48
DANIEL ALAN BRUBAKER
“and;” the lam- ’alif has been omitted. The meaning
here is thus “you and your fathers” rather than the
1924 Cairo edition’s “you nor your fathers”
What reads mubdrakun musaddiqu, “[it is] blessed
and confirms,” in the 1924 Cairo edition of Q6:92 is
written in this manuscript without the long medial
alif in the first word and also with a long 7alif at the
end of both words, to render mubarakan
musaddiqan, apparently “a blessed and confirming
one.”
The waw “and” that precedes li-tundhir, “that you
may warn,” in the 1924 Cairo edition is absent on
this page.
Whatreads salatihim (archigraphemically CLA
BHM), “prayers,” in the 1924 Cairo standard is
written in this manuscript with a waw instead of
the medial long “alif, that is, salawatihim, '* or,
archigraphemically, CLW BHM. The latter is plural;
a slight change of meaning.
The ay, “or,” of Q6:93 is written in this manuscript
as wa, “and,” to render “he who imputes falsehood
to Allah and says,” instead of the 1924 Cairo
edition’s “he who imputes falsehood to Allah or
says.”
The 1924 Cairo standard’s idh, “while,” of Q6:93 is
written in this page as idha, “when.”
The long 7alif that is in second position of the bastu,
“outstretched,” of Q6:93 in the 1924 Cairo edition is
missing on this page.
The word aSy rabbikum, “your Lord,” is written
between allah and fa- ’innd of Q6:95. This does not
Corrections in Early Qur’an Manuscripts 49
exist in the 1924 Cairo edition but does make
grammatical sense here, reading, “That is Allah
your Lord, how then,” rather than “That is Allah,
how then’ as it exists in today’s standard. Itis
interesting that the correctors of this page did not
erase this word. Did they feel it belonged here?
The points above give a sense of the variant character of
this manuscript. Now we will discuss the corrections on this
page. There are at least five:
1) THERE IS an erasure near the end of line 3, of two words
whose shadow partly remains. It occurs after the hawlaha wa,
“around it, and,” of Q6:92 and before the alladhina, “those
who,” that follows it. A significant gap remains. The result at
this point is a rasm that conforms to the 1924 Cairo standard.
2) ON THE 6th line pictured, the word «ile ‘alayhi, “against
him,” has been written over an erasure in Q6:93 following the
words bima kuntum taqulun, “for what you (pl.) used to say.”
However, the 1924 Cairo edition does not read ‘alayhi here,
and moving closer to conformity might be the reason for what
I believe to have been the next intervention, noted in point #3.
3) A SUBSEQUENT correction was made at Q6:93, this time in
the right margin, where 4! gle ‘ala allah, “about Allah,” has
been written, but oddly without erasing the ‘alayhi that it is
apparently intended to supplant. Also, this phrase is written
50 DANIEL ALAN BRUBAKER
next to the start of the following line, but it seems to be
intended for this spot.
The more interesting thing here is that the page at this line
remains out of conformity with the 1924 Cairo edition in that it
includes the additional words 5 tb (y9)4S takfurtna bi-llahi wa,
“they disbelieve in Allah and,” after the words pi La: bima
kuntum and (91% taqulun of this verse. That those words were
not only plainly written in this manuscript here at the time of
its production, but also allowed to remain in it after two
rounds of correction, despite not being part of the 1924 Cairo
edition, seems important.
4) ON THE 8th line pictured, the word (iI alladhina, “whom,”
of Q6:94 has been inserted where it was at first omitted.
5) AT THE beginning of the second to last line of the page, the
word ¢y9alas ya amuna, “they know,” of Q6:97 has been written
over an erasure. The shadow of what was first written can still
be seen and its archigrapheme appears to be BHMW N;
however, this archigrapheme renders no word in the Qur’an.
It is possible, I suppose, that the erased text was BEMHW N;
this could correspond to one word, (y9¢«2: ya ‘mahun,
“blind/dumbfounded,” a word that occurs only seven times in
the Qur’an, with one of them being at the end of Q6:110, that
is, in close proximity to this verse. If this were indeed what was
first written here, the verse would have read “We have made
plain the signs for a people who are blind.” It is difficult, at this
point, to make a strong opinion on this, since the new writing
Corrections in Early Qur’an Manuscripts 51
covers the erasure partially. So, while possible, it is not at all
clear that an ‘ayin was present.
overwritten ARE i bi Teas
Chass ——
FIGURE 18: Q6:92-97 in the 1924 Cairo text, with the MS.474.2003
corrections shown
52 DANIEL ALAN BRUBAKER
Example 6: Multiple post-production corrections in a
1st/7th century Qur’4n
FIGURE 19: MS.67.2007.1 (Brubaker photo, by permission of the Museum of
Islamic Art)
THIS FRAGMENT, and two others grouped under sequential
shelf numbers, is of a similar time period and style to the
Codex Petropolitanus (BnF arabe 328a-b, etc.) and also the
Birmingham folios that Alba Fedeli brought to the world’s
attention, which were radiocarbon dated with a very early date
range, a 95.4% probability of the animal having been last alive
between 568-645 AD.!3 Several years prior to the testing of the
Birmingham folios, parchment from the palimpsest San‘a’ I
were similarly radiocarbon dated, giving a 95% confidence
date range of 578-669 AD.'4 BnF arabe 328 is a vertical bifolio
written in Ma’ il/hijazi script.
The corrections here are found in MS.67.2007.1, in the
Museum of Islamic Art in Doha. Inserted are the words wa-
‘amilu 7I-sélihati thumma ttaqaw wa-°’amani of Q5:93. The
main insertion has been made above the main line of text.
Except possibly for the first portion, wa- ‘amili, about which I
have some question because of the way it is written, this inser-
tion looks to be the work of the original scribe and was prob-
ably made soon after the time of first writing. Q5:93 has several
Corrections in Early Qur’an Manuscripts 53
repetitions in it, and it is not at all surprising that a scribe
might have become confused and made a mistake that needed
to be corrected later. This correction is thus almost certainly
due to a simple copyist error at first writing.
There is one part of this insertion, however, that appears to
be part of a second and later correction. It is the final ’alif of
‘amilu, “they did,” and this orthography of the third person
plural ending, I think, was omitted at first correction and
added later. Also, the corresponding 7alif of ‘amanu, “they
believed,” at the very end of this insertion is missing, a further
odd detail, given that it is typically used elsewhere on this page
and was also added in at the end of ‘amilu.
Finally, and perhaps most interestingly, the initial ’alif of
Igiua! ’ahsanu, “they did good (imperative, 3rd pl.),” was
omitted at the time of first writing and was added later, but
with red ink, the same ink used for the dots representing short
vowels elsewhere on this page.
So, there is a lot going on in some of these manuscripts,
and a close and careful examination is needed. I almost did
not notice the issue with the ’alif of ’ahsant myself. There
have been several times (I remember one quite clearly at the
Bodleian Library in Oxford several years ago), when I have
been working closely and carefully with a page for a long time
and have almost been ready to move on before noticing a
correction that ought to have been plainly obvious much
sooner. It is a reminder that patience, humility, and attention
to detail is essential in this work.
54 DANIEL ALAN BRUBAKER
hse <a “Atl
2 ee Files ge tibanes
naan poco insertion
ul "ee
FIGURE 20: Q5:93-94 in the 1924 Cairo text, with the MS.67.2001 corrections
shown
The complexity of the above situation may be taken as
evidence that this manuscript was in use and was felt impor-
tant enough to correct, multiple times. The issue with the red
‘alif is interesting because it is not a matter of orthography or
reading. So, there is more work remaining to be done on this
section and in this manuscript, which has layers of informa-
tion to unpack.
Corrections in Early Qur’an Manuscripts 55
Example 7: Post-production insertion of the words “the
seven”
FIGURE 21: BnF arabe 327, fol. rr.
BnF arabe 327, in the National Library of France, is written
in Déroche’s B.Ib style. A fragment of 14 folios, its pages are
nearly square, 26-27 cm. (about 10.5 inches) tall and just
slightly wider, with 18 lines of writing per page. It probably
dates to the 8th century. I have noted nine different corrections
in this manuscript, and | believe one of these was corrected
more than once.
Two different corrections can be seen in Figure 21. The first
is above the upper line shown, where the words gusu!! al-sab %,
“the seven,” of Q23:86 have been added by a later scribe and in
a very different script style from that of the original scribe. As
first written, this portion read, “Say: ‘Who is the Lord of the
heavens and the Lord of the Great Throne?”” As corrected, and
as it stands in the 1924 Cairo edition, it reads, “Say: ‘Who is the
Lord of the seven heavens and the Lord of the Great Throne?”
56 DANIEL ALAN BRUBAKER
Clearly, the verse makes sense either with or without the word;
the only question is which reading reflects the original.
The number seven occurs several places in the Qur’an, but
is not a strong motif as it is, for example, inside the Bible.
There is one other folio, probably 8th century but possibly late
7th, from among the San‘a’ manuscripts that omits the word
seven in Qg:80 where the word does exist in the standard text
today.'!° That page, although corrected elsewhere, is not
"16 was allowed by the
corrected at this point; its “omission
corrector to remain. The details in that manuscript, at a verse
that seems to have strong intertextual connotations, around
the number seven have led me to suspect that there was some-
thing going on relating to this word “seven.” Qg:80 is talking
about forgiveness, and, with the inclusion of “seventy,” it
suddenly shares two elements with Matthew 18:21-22, “Then
Peter came and said to Him, ‘Lord, how often shall my brother
sin against me and I forgive him? Up to seven times?’ Jesus said
to him, ‘I do not say to you, up to seven times, but up to
>
seventy times seven.” To be clear, the omission at Qg9:80
occurs in a different manuscript from the one pictured above,
but, because of it, I find any insertion or a variant involving
“seven” or “seventy” to be interesting.
The lower line pictured above also has a correction, by a
different corrector, I think. It is an inserted “alif in front of li-
Ilahi “Allah’s” of Q23:87. The result does not align with the 1924
text, but it does comport with Abu ‘Amrs reading (and anoth-
er); it is standard in some parts of the world today. The effect is
to convert the word “Allah’s” to “Allah.” This word is an answer
to the question posed in the prior verse, “Who is the Lord of
the seven heavens and the Lord of the Great Throne?” This
conversion, at this particular verse, has been discussed by
Corrections in Early Qur’an Manuscripts 57
Cook, who notes the resultant reading as allegedly aligning
with the codex sent by ’Uthman to Basra, as described in al-
Dani, who ascribes the insertion to Hajjaj.!’
An oval mark, partially pictured, also lies over the words
following this correction. It marks that the words ,4)¥ ly wa-l-
’ard, “and the earth” (which is not present in the 1924 text) for
omission, to be replaced by the inserted al-sab %.!® Thus, “the
heavens and the earth” has in this manuscript become “the
seven heavens.”
pane
r
pe & adel sal Hels
et ae EG AB]
Qe ad nies Ojdicee vale
FIGURE 22: Illustration of location of correction Example 7 compared with
the 1924 Cairo text
58 DANIEL ALAN BRUBAKER
Example 8: Erasure overwritten in a 1st/7th century Qur’an,
possibly by original scribe and likely soon after original
production
"y Af
javae 5
" ad
Pe oe | wi kop . P Agee
a = wy 4 . 44 é i
ot
Pe, Le
olis aol a
FIGURE 23: BnF arabe 330, fol. 55r.
BNF ArABE 330 is a fragment of 69 large vertical parchment
folios, approximately 37 cm. (14.5 in.) tall by 28 cm. (11 in.) wide.
It is a composite manuscript; its folios are not all from the
same original Qur’an. Prof. Déroche classified its folios under
various script styles: hijazi III, A.I, and B.Ib.!? He has recently
classified the portion 330c as style O I,”° and considers it to be
part of the Fustat Umayyad Codex.*! The page shown (from
330g) above remains, for the moment, unclassified.?” I have
visited BnF arabe 330 twice and have noted 65 corrections
among its pages.
In the example pictured above, allah (“Allah”) of Q4:149 has
apparently been replaced by 4S «il! allahu kana, “Allah is,” via
erasure and overwriting. This change appears to be the work
of the original scribe and may have happened as part of the
production process (for example, after proofreading the line or
the entire page when it was first written). Probably the word
Corrections in Early Qur’an Manuscripts 59
kana was initially omitted, since the phrase fa-’inna Ilaha
‘afuwwan gadiran, “so surely Allah is Forgiving, Powerful,” and
was corrected to remedy the faulty grammar.
The verse carries the same sense with or without this
word, but its inclusion is standard today. I have found no
mention of an issue at this spot in the gird at literature.
This is not the only correction on this folio; seven lines
below it there is another erasure that has been overwritten.
Sys 432) @ Hse
.r VE, Ahn) 54 ales Gilg ol reli dps 2A 54) ri
‘ \
2?2| te é i 2) ie eg 4 A Con fd. © 28
(8 by Eg and I ps sae Feld PE
“T pee 4 a aati P fod < -
| Batti basic ih ak ih haa »
Me sds sdb chal, peat +i feel ;
FIGURE 24: Illustration of location of correction Example 8 compared with
the 1924 Cairo text
60 DANIEL ALAN BRUBAKER
Example 9: Post-production insertion of “the Merciful”
ry aw Jt <x.
FIGURE 25: BnF arabe 327, fol. 12v.
This is the second example of correction from BnF arabe
327. A general description of the manuscript can be found in
Example 7 above.
In this case the words acu! al-rahim, “the Merciful,” of
Q42:5 were omitted at the time of production and have been
added in above the line at a later time. The correction in this
case appears to be the work of a different scribe. It is the last
word of the verse, and completes the pair of attributes of Allah
that commonly ends a verse. As first written, the verse read,
“and Allah is the Forgiving.” As corrected, and as is standard
today, it reads, “and Allah is the Forgiving, the Merciful.”
There are two additional interesting things about this
correction. First, it looks like it has been written with two
different nibs, one very narrow and the other a little wider,
though still not as wide as that which was used for the original
production of this page. Second, the correction itself appears
to have been rubbed out or almost erased at some point.
Corrections in Early Qur’an Manuscripts 61
The verse is grammatically correct and semantically viable
without the insertion, but its absence throws off the standard
rhythm, as the expectation is usually for a pair of attributes of
Allah at the end ofa verse. Also, the first word of the pair, y544J/
al-ghafur, “the Forgiving,” does not fit the rhyme pattern of the
other verse endings in this chapter, while the inserted al-rahim
does. There are a number of places in the Qur’an where verse
endings do depart from the overall rhyme pattern, and a devia-
tion can serve a poetic purpose,” but it is difficult to imagine
reading this verse with only the single attribute.
AAa____
a
A@ ssc ait! NS
sacle asl
ee
pa
OSs pee eit,
FIGURE 26: Illustration of location of correction Example 9 compared with
the 1924 Cairo text
62 DANIEL ALAN BRUBAKER
Example 10: Post-production mid-line insertion in a 1st/7th
fae
BnF arabe 331 is also in the National Library of France. It is
century hijazi manuscript
FIGURE 27: BnF arabe 331, fol. rv.
a well-preserved fragment of 56 very large parchment folios,
about 39.5 cm. (15.5 inches) tall by 34 cm. (13.5 inches) wide. It
has about 19 lines of writing per page. Its script style has been
identified by Déroche as B.Ia.”4
In the detail of this manuscript shown above, the word Jie
”
mithli, “as,” of Q2:137 was omitted at first writing and then
added in at a later time, with the preceding bi. The correction
is ina very different hand using a much narrower nib; it looks
almost like a modern intervention on the page. This having
been said, the ink used for the insertion is very close in color
and consistency to that of the original writing on this page.
Probably this was just a very good match in ink, but it bears
mention.
One interesting feature of this case is that the bi which was
first written, linking forward to md, has not been erased, so as
Corrections in Early Qur’an Manuscripts 63
it stands now it has an extra letter when compared to the 1924
Cairo text, with the portion written as Las Jia: Isis! aman bi-
mithli bi-ma, an apparently non-viable reading.
As this page was first written, the verse makes grammatical
and semantic sense, “If they believe in that which you have
believed,” versus the 1924 Cairo text, which is approximately
translated, “If they believe similarly to that which you have
believed.”
Ss ae tir fi cp ok WS
kas bieeosh oa Ye sit)
9 V8) 28 s2t\4 J las
| extra denticle remains
tae
4st; a :
FIGURE 28: Illustration of location of correction Example 10 compared with
the 1924 Cairo text
64 DANIEL ALAN BRUBAKER
Example 11: Post-production marginal insertion of “Allah”
in the Topkapi codex
i thy
" Midd
FIGURE 29: Topkap1 codex, fol. 374v. (Source: Altikulag, Tayyar, Ed. Al-
Mushafal-Sharif attributed to ‘Uthman bin ‘Affan (The copy at the
Topkap1 Palace Museum). Istanbul: IRCICA, 2007.)
This is a second instance of correction from the Topkap1
codex, and also a further insertion involving the word allah.
This insertion of lam-lam-he occurs near the beginning of
Q66:8. As originally written, the first allah of this verse was not
present. This change has been made with a very small nib and
probably occurred long after the first production of this
manuscript. It is possible this addition is a modern inter-
vention.
Prior to the insertion, this could have been read “Oh you
who believe! Turn to a sincere repentance,” were it not for the
Corrections in Early Qur’an Manuscripts 65
original inclusion of the alif after *ila. There is obviously a
certain range of possibilities for some of these archigraphemes
— for example, if some letters were to be pointed differently
than they are today, rendering different consonants — that
could open alternate readings, but the one with the most flexi-
bility when undifferentiated has a dot underneath it in this
manuscript, binding us to ba’. So, it is not clear to me what
was intended by the original version, or whether it could have
been read viably. It is worth noting, however, that the word in
question is also part of a section that has been erased and
overwritten in the manuscript Marcel 104 in the National
Library of Russia, and that correction will be featured, among
many others, in my larger forthcoming book.
\ ar zal ip aes 1 Ve %
Np oe ae te
FIGURE 3o: Illustration of location of correction Example 11 compared with
the 1924 Cairo text
66 DANIEL ALAN BRUBAKER
Example 12: Erasure overwritten and stretched in a Ist/7th
century Qur’an
FIGURE 31: BnF arabe 328, fol. 8r.
This example shows another correction in BnF arabe 328a,
part of the Codex Parisino-Petropolitanus. This manuscript
was already introduced in Example 2.
The correction shown here is found on folio 8r, near the
beginning of line 13. In it, the dad-lam of 4.45 fadlin, “bounty,”
of Q3:171 has been written over an erasure. Erasure marks are
clearly seen, including some of the shapes of the letters of
what was first written here, among which were four upward-
extending letters, the first of which is preceded by a short
tooth letter. The corrector has used a different nib and ink
from that which was used in first production of the page; also,
the hand and angle of the script vary from that used on the
rest of the page. This change is clearly a later intervention. The
length of the space that is now covered by these two letters is
5.3 cm, and this would typically be occupied by five to eleven
letters elsewhere on the page. There is only one other place on
this page where this much space contains as few as five letters.
Corrections in Early Qur’an Manuscripts 67
The result of this correction is a rasm that conforms at this
point with that of the 1924 Cairo edition.
uw Gees areas ail Jae Fa
Age
FIGURE 32; Illustration of location of correction Example 12 compared with
the 1924 Cairo text
I have observed particular verses” and words”® that are
frequently corrected in Qur’an manuscripts. The word fadl is
not frequently corrected, but it is of theological significance.”/
The word 41.44 fadaina, “we have favored,” of Q6:86 has been
written over a covering in the Cairo mushaf al-sharif. This is the
only correction so far that I have noted for any part of Q3:171.
68 DANIEL ALAN BRUBAKER
Example 13: Erasure overwritten, apparently changing the
verb declension
oy yo “a
+ ita 9: a,
wer 9
FIGURE 33: BnF arabe 340, fol. 26r.
BnF arabe 340 has 121 folios that are written generally in
horizontal format on parchment. It is another composite frag-
ment, meaning that its folios did not come from the same orig-
inal book, but rather several. A number of its pages are in later
script styles of the roth and even 1th centuries (e.g. D and
NS).28 Déroche dates arabe 340(f) to the 9th century.” A
number of these folios has been classified B.IJ,2° which would
be gth century. Folios 1-12 and 13-30 (which includes the folio
shown here) he has left unclassified.*! The folio shown here,
from BnF arabe 340(b), is probably early 9th or even 8th
century.
Even though they represent different original codices, it
will give readers a sense of scale to know that I have noted 91
corrections in the pages of BnF arabe 340.
The example above is found on 26r, one of the folios whose
script style Déroche has left unclassified, in the middle of line
2. Here, the final lam of Jl qala, “he said,” has been erased, and
in its place lam-wdw-alif has been written. The result is the
word Isls qalu, “they (m. pl.),” of Q34:35. As it was first written,
this verse read, “And he said, ‘We are more [than you] in
Corrections in Early Qur’an Manuscripts 69
2
wealth and in children.” As it stands now on the manuscript
page, and as it exists in the 1924 Cairo edition, this verse reads,
“And they said, ‘We are more [than you] in wealth and in
children.”
This is not an extremely dramatic correction, and there are
others among the pages of this fragment that are actually more
interesting, but my purpose in this book is not to pick the most
dramatic corrections but rather to show the range of the
phenomenon. Conversions involving Jl, or variations on this
theme (in this case the third person plural), are among the
most common types of correction in early Qur’ans.
2
~ 4264
x Sec abu sO Site
FIGURE 34: Illustration of location of correction Example 13 compared with
the 1924 Cairo text
70 DANIEL ALAN BRUBAKER
Example 14: Erasure leaving a gap in Topkap1 codex
FIGURE 35: Topkapt codex, fol. 657, showing erasure of a single *alif at the
end of line 11. (Source: Altikulag, Tayyar, Ed. AJ-Mushaf al-Sharif
attributed to ‘Uthman bin ‘Affan (The copy at the Topkap1 Palace
Museum). Istanbul: IRCICA, 2007.)
a pe Jew
at tm
- y ;
, ey
FIGURE 36: Topkapi codex, fol. 65r, showing erasure of \lah qad at
beginning of line 12. (Source: Altikulag, Tayyar, Ed. Al-Mushaf al-Sharif
attributed to ‘Uthman bin ‘Affan (The copy at the Topkapi Palace
Museum). Istanbul: IRCICA, 2007.)
Here is a third example from the Topkap: codex. In this
Corrections in Early Qur’an Manuscripts 71
instance, there has been an erasure of two words at the begin-
ning of line 12, with the first letter of Allah having also been
erased at the end of line 11. The shadow of what was first
written remains; it is s6 4! allahi gad, “Allah has already,” of
Q4:167.
This verse is grammatically and semantically viable
without the portion that has been erased. Whereas before the
correction the consonantal text of this portion read, “Surely
those who disbelieve and hinder from the way of Allah have
strayed far into error,” after the correction it reads, “Surely
those who disbelieve and hinder from the way have strayed far
into error.”
The reason for this erasure is unclear, but its precision in
taking out only the selected words is evident. Someone, at the
time of this correction, evidently thought that those words did
not belong in this place.
This is an unusual correction, as it takes the page at this
point away from conformity with the 1924 Cairo text. Such
changes represent a very small proportion of the total number
of corrections I have noted; usually corrections result in a rasm
that conforms, or that conforms more closely than it did
before, with what is standard today. It is natural and reason-
able to presume that a corrector felt the change that he was
making to the page to be movement toward a more correct
text. So in cases like these — and this one in particular is an
excellent example — the question of what formed the basis
for such a belief on the part of the corrector is intriguing.
It does look from the facsimile as though there has also
been an erasure on line 10, just before the end of the line. I'll
not describe it here, partly because I am not sure about it. It’s
always best to look at the manuscripts directly; even an
72 DANIEL ALAN BRUBAKER
extremely good photograph does not measure up to direct
examination. Of course, as a practical matter, these objects
cannot be handled by everyone, so when I go to look at them I
look very closely and make detailed notes.
tps Ob ae
ae ANC i gies
FIGURE 37: Illustration of location of correction Example 14 compared with
the 1924 Cairo text
Corrections in Early Qur’an Manuscripts 73
Example 15: An erasure leaving a gap in an 8th or 9th
century manuscript
FIGURE 38: MIA.2013.19.2, verso. (Brubaker photo, by permission of the
Museum of Islamic Art)
THIS Is a fragmentary partial folio on parchment in the
Museum of Islamic Art in Doha, Qatar. Its writing is very
similar to that of the Topkapi codex and its style is C.Ib.
In this example, there is an erasure at the end of one line
and the beginning of the next. It occurs after the word 4L.as
fadlihi, “his grace,” of Q24:33. The next word after the erasure is
the word that follows fadlihi in the 1924 Cairo edition, that is,
eval, wa-lladhina, “and those who.” What was first written in
this space that is now empty cannot be discerned from the
manuscript, as no shadow indicating the shape of the letters
remains.
I have noted two manuscripts that have multi-word correc-
tions of this verse; the other is BnF arabe 327, in which a long
portion of text has been written over an erasure and appar-
ently corrected more than once. That is an interesting correc-
tion, and I will certainly talk more about it in a subsequent
publication. However, that correction does not cover the
section of the verse at issue here; they do not overlap.
74 DANIEL ALAN BRUBAKER
oe ON JO@ _ «tlie Sh -p
FIGURE 39: Illustration of location of correction Example 15 compared with
the 1924 Cairo text
THE RASM of this page as it now stands aligns at this spot with
that of the 1924 Cairo edition, but as this page was first written,
it contained something extra. Since this manuscript is the only
known copy with a correction at this point, we must wait to see
if anything else emerges in future research. Perhaps there was
a mere scribal error.
Corrections in Early Qur’an Manuscripts 75
Example 16: A post-production insertion in the Cairo
Qur’an
FIGURE 4o: Cairo mushaf al-sharif, fol. rogr. (Source: Altikulac, Tayyar,
Ed. Al-Mushaf al-Sharif attributed to Uthman bin Affan (The copy at
al-Mashhad al-Husayni in Cairo). (2 vols.) Istanbul: IRCICA, 2009.)
This example comes from the monumental codex that is
kept on display in the Husayni Mosque in Cairo. This codex is
an enormous book of 1088 parchment folios. Like the Topkapi
mushaf already mentioned, it has been popularly claimed by
its custodians and the governing authorities to be one of the
mushafs of the Caliph ‘Uthman. This opinion is rejected by
scholars, including Dr. Altikulas, who places the time of
production of the Cairo mushaf at the end of the 8th or begin-
ning of the 9th century.** I will give more detail about this
manuscript at the end of this chapter.
In this case, the (,lS kana, “is,” of Q4:33 was not written in
this verse at the time of production of this manuscript.
Though only the first two letters of this inserted word are now
visible in this facsimile photograph, the full word kana was
presumably added here, with a very fine nib. I would like to
76 DANIEL ALAN BRUBAKER
have the opportunity to look at this manuscript directly to
confirm this theory. The verse makes sense with or without the
word, and its meaning is about the same in either case: “And
Allah has power (lit. ‘is powerful’) over all things.” As in many
semitic languages, the verb “to be” is often not used as its
sense is implied when the adjective directly following the
noun it modifies. The verb can be included, but it is not gram-
matically necessary.
He ales ae Z Sh ee
oot i Olah | zah tate at
laf Agsthee Lehi
a 7" * | inserted
e 2
ox
FIGURE qr: Illustration of location of correction Example 16 compared with
the 1924 Cairo text
This manuscript is not the only one in which the word
kana has been inserted. A similar insertion of kdna exists in
NLR Marcel 17, folio tv, at Q4:6; however, in that case, it does
not appear that as much time elapsed between original
production and correction.
Corrections in Early Qur’an Manuscripts 77
Example 17: A correction involving “allah”
FIGURE 42: NLR Marcel 11, fol. 7r.
This erasure has been overwritten in Marcel 11, part of the
so-called Fustat Umayyad codex, which has already been
introduced under Example 3 above. Written in script style O I,
this is a vertical fragment of 12 folios measuring about 36.5 cm.
(~14.5 in.) tall by 31 cm. (~12 in.) wide. It has 25 lines of writing
per page, and its folios are quite delicate now. This particular
fragment has a very high density of corrections: I’ve noted 46
of them over its 12 folios. Furthermore, some of its corrections
are quite dramatic.
This correction is found on 7r, in the middle of line 9. All
but the first two letters of «ii! 423 ni‘mata Ilah, “the favor of
Allah,” of Q33:9 has been written over an erasure. A different
nib and ink have been used, and the writing is that of a
different hand. Also, the writing has been bunched in. My
impression is that this could have read ni ‘matihi, “his favor,” at
first; this rendering would have fit the space and would make
grammatical sense here. However, this interpretation is just a
reasonable conjecture; I cannot tell for sure.
78 DANIEL ALAN BRUBAKER
We
SV GN) GM GAA gg VM SAA I
Le ey Peel
written over erasure
ipo ea cd
I Sah daly
Kel Gl 57 fs sice agli aibse &
Ki, Jishys eats AIO Piaget
dibs N55 pt Alo, Meitlis 5 ai cet PE
Iss eal i L168 © 11
FIGURE 43: Illustration of location of correction Example 17 compared with
the 1924 Cairo text
In addition to its many other corrections and this one, the
12 folios of Marcel 11 have four omissions of allah that were
later inserted: 33:18, 33:24, 33:73, and 41:21 — these were shown
in Example 3.
Surah 33 has a fair number of corrections in the early
manuscripts. Most of them are fairly small and many involve
orthography. There is a more lengthy erasure overwritten at
verse 73 in BnF arabe 340. I will discuss this and others further
at a later time.
Corrections in Early Qur’an Manuscripts 79
Example 18: Post-production insertion of “the hour” in a
3rd/9th-4th/10th century manuscript
FIGURE 44: NLR Marcel 7, fol. 7r. (Brubaker photo, by permission of the
National Library of Russia)
Marcel 7 is a horizontal parchment Qur’an fragment of 10
folios. Its pages measure 17.7 cm. (~7 in.) tall by 23.3 cm. (~9 in.)
wide. Its script style is probably D.IV and it is thus likely a 9th
or 10th century manuscript. I have noted 8 corrections across
its 10 folios, a high frequency of correction for a manuscript
produced more than two centuries after the time of ‘Uthman.
In this case, the word 4¢Lull al-sa ‘ah, “the hour,” of Q6:40
has been written in as a superscript insertion. It has been done
with a very narrow nib and in a different hand. It is possible
that this isa more modern correction.
This word, al-sad ‘ah, has been corrected in other manu-
scripts. In BnF arabe 340, there is a rather lengthy erasure over-
written at QI5:85 that includes the instance of this word in that
verse; it is not at all clear, however, that in this case the correc-
tion had to do with this word particularly or with another. The
word sda ‘ah is also written over an erasure at Q7:34 in the
80 DANIEL ALAN BRUBAKER
manuscript E20, located at the Institute of Oriental Manu-
scripts, also in St. Petersburg.
<A si e_ WHO Fi,
eda a es 36 al
FIGURE 45: Illustration of location of correction Example 18 compared with
the 1924 Cairo text
Corrections involving “the hour” are interesting since this
word relates to eschatological (that is, referring to end times)
or apocalyptic themes, which an insertion such as this one
would have strengthened and clarified.*? In both 6:40 and
15:85, the word al-sa‘ah are eschatological references to the
hour.
Corrections in Early Qur’an Manuscripts 81
Example 19: Erasure overwritten involving “allah”
y
th a 1+
FIGURE 46: NLR Marcel 5, fol. 11r. (Brubaker photo, by permission of the
National Library of Russia)
A
Housed in the National Library of Russia, Marcel 5 is a
parchment fragment of 17 folios from a large format Qur’an.
Its pages measure 50 cm. (~19.5 in.) tall by 35 cm. (~14 in.) wide.
Its text block measures 44 x 30 cm. (~17 x 12 in.). It has 20 lines
of writing per page. Many of the letters are differentiated by
diacritics, which are present in the form of fine diagonal nib
marks; these seem to be original to the manuscript.
The above correction is found on folio 1 recto. In it, the
words 4] 58 huwa Ilah, “he is Allah,” of Q34:27 have been
written over an erasure. This is not the work of the original
scribe; the ink is different and the letters are drawn in rather
than written. The huwa, which extends into the right margin,
Corrections in Early Qur’an Manuscripts 83
Example 20: Erasure overwritten of nearly a full line of
text, involving “provision”
FIGURE 48: MIA.2014.491, fol. 7v. (Brubaker photo, by permission of the
Museum of Islamic Art)
Located in the Museum of Islamic Art in Doha, this object
is a small horizontal format bound Qur’an fragment of nine
parchment folios, measuring approximately 17.5 cm. (~7 in.)
tall by 28 cm. (~11 in.) wide. Its script style is B.II.
This fragment contains several interesting corrections.
Shown here is the erasure and overwriting of an entire line in
the middle of folio 7. The new text is (8%) agidy, Las 3 wa-
mimma razaqnahum yunfiquna, “and of that which We have
provided them,” of Q8:3, plus the initial ’alif of the following
verse. Erasure marks are quite clear on this page and the
current writing on this line is somewhat stretched out to fill
the space, an indication that what was first written here was
longer.
The word rizq, “provision,” is directly corrected or part of
larger corrections (as is the case here) quite frequently in early
Qur’ans. It was such a prominent feature that it topped my list
of frequently corrected words in early Qur’an manuscripts in
a conference paper I delivered at the International Qur’anic
Studies Association several years ago. I am not yet sure why
rizq is so frequently corrected, that is, what the issue is, but I
would not be surprised if the issue has played into the motiva-
tion for this particular instance of correction.
84 DANIEL ALAN BRUBAKER
itten over Brenune
FIGURE 4g: Illustration of location of correction Example 20 compared with
the 1924 Cairo text
This concludes my presentation of the twenty examples. I
am aware that, with Examples 3 and 5 (in particular), which
contain several corrections each, I have actually shown more
than twenty. My intent was to be generous, in the spirit of the
old American tradition of the “bakers’ dozen.” Also, I wanted
to take the opportunity to demonstrate some apparent
patterns of correction (such as that in Example 3) that would
be more difficult to see if ] merely described them separately.
~~
Corrections in Early Qur’an Manuscripts 85
Another phenomenon: Covering in the Cairo Mushaf?
FIGURE 50: Cairo Mushaf al-Sharif, fol. 33v. (Source: Altikula¢, Tayyar, Ed.
Al-Mushaf al-Sharif attributed to Uthman bin Affan (The copy at al-
Mashhad al-Husayni in Cairo). (2 vols.) Istanbul: IRCICA, 2009.)
WHEN SURVEYING manuscripts for my doctoral dissertation, I
came across some instances of what looked to me like writing
that had been covered up. In an abundance of caution, I did
not classify them as corrections, and even now am reluctant to
do so since I have not had the opportunity in most cases to
86 DANIEL ALAN BRUBAKER
look at the manuscripts in question directly to make a very
careful assessment.
Pictured above is one page of the monumental Qur’an, in
the opinion of Altikulag probably dating to the end of the 8th
century or the beginning of the 9th. It is an interesting
manuscript for a variety of reasons, not least of which is its
movement between conformity with one or another of the
various documented codices:
The comparison we made between the Mushafs attributed
to Caliph ‘Uthman in 44 places concerning pronunciation, a
superfluous or a missing letter and the structure of words
leads us to think that this Mushaf is not related to any of the
Mushafs of Caliph ‘Uthman. [...T]his Mushaf differs from
the Medina Mushaf in 14 of the 44 places, from the Mecca
Mushaf in 15 places, from the Kufa Mushaf in 7 places, from
the Basra Mushaf in 9 places and from the Damascus
Mushaf in 28 places. As a result, although the Cairo Mushaf
has common points with one or more than any one of these
Mushafs in each of the 44 places, it is not exactly the same as
any one of them.*4
This manuscript has more than 1,000 folios. Many of them
have similar tapings that cover portions of the text. In my
experience, such tape is sometimes used to repair a weak spot
on the page, such as where the acidity of the ink has eaten
through the parchment over the centuries, and I have
observed at least one instance of such tape being applied for
the purpose of repair in a manuscript fragment of a similar age
and script style to the Cairo mushaf. Indeed, on many pages of
the Cairo mushaf, parts of what is written beneath the tape
Corrections in Early Qur’an Manuscripts 87
extend beyond the tape edges and appear to be in alignment
with what we would expect to be there when compared with
the 1924 Cairo edition.
My first objective with tapings like the one above is there-
fore to rule out the possibility that the tape was applied merely
for the purpose of page repair. Were the manuscript in front of
us, we could look at the page to assess its condition, and also
examine the back of the page to see if there is evidence of split-
ting or weakening at the spot where the tape has been applied
on the reverse.
In the case of the Cairo mushaf, I've not yet been able to
survey these pages in person. I hope to be permitted to do so
one day.
In absence of the opportunity for direct inspection, then,
we must work from photographs, and the first thing I do
after looking closely at the side with the tape is to look
closely at the photograph of the reverse side of the same
page. In many instances in the Cairo mushaf, as in folio 33
pictured above, the reverse of the page appears to be
perfectly sound. This observation leaves open the possibility,
then, that the tape might be serving another purpose, such
as selective concealing of something that is written on the
page.
If the rasm beneath the tape of the page shown above
conforms to the 1924 Cairo edition, then the covered portions
would be as follows:
e Line1- All but the first three letters of (4. apa als
éus wa-?akhrijiham min haythu, “drive them out
from wherever,” of Q2:191
e Line 5 - All but the first two and last two letters of
88 DANIEL ALAN BRUBAKER
aSshs (Ls fa- *in-qatalakum, “so if you fight to kill
them,” of Q2:191
e Line 6- All but the first five letters of ell3S aaglisls fa-
qtuluhum kadhalika, “then kill them (imper.), such,”
of Q2:191
e Line7- All but the last five letters of !5g34! cls fa- ’ini-
a 3
ntahat, “and if they desist,” of Q2:192
e Line 8- The first three letters of y94¢ ghafurun,
“forgiving,” and the last three letters of ass,
rahimun, “merciful,” of Q2:192
e Line jo - All but the first letter of at ¢4!! al-dinu li-
Ilah, “the religion belongs to Allah,” of Q2:193
e Line 1 - All but the last letter of (y9se ‘udwan,
“enmity,” of Q2:193
e Line 12 - The final two letters of .g.i!L bi-?I-shahr,
“in the month,” of Q2:193
Until I can see what lies under the tape, I do not know
what has been covered up in each case. Still, I think it is worth
mentioning that these coverings exist, and in many cases seem
to have been applied when there was no need of page repair,
possibly to hide what was written on the page at particular
points.
Corrections in Early Qur’an Manuscripts 89
FIGURE 51: Cairo Mushaf al-Sharif, fol. 430r. (Source: Altikulag, Tayyar,
Ed. Al-Mushaf al-Sharif attributed to Uthman bin Affan (The copy at
al-Mashhad al-Husayni in Cairo). (2 vols.) Istanbul: IRCICA, 2009.)
FINALLY, there is the matter of coverings overwritten. Above is
one example of this. There are many places in the Cairo
mushaf where these tapings have been written over. In the
photo above, this appears to have happened in three places:
e On the first line pictured, all but the first two letters
of aguail bi-’anfusihim, “in themselves,” of Q13:I1
has been written on the top of such a taping.
e On the second-to-last line pictured, all but the
initial ’alif of aS (sil! alladhi yurikum, “he who
shows you,” of Q13:12 has similarly been written
over a taping, and is rather stretched out. The
stretching is not unusual in this manuscript, but it
is more pronounced in this spot than is standard
for the original scribe. It is notable that the way this
section is written over the tape is missing one letter
go
DANIEL ALAN BRUBAKER
when compared with the 1924 Cairo edition, which
has an additional ya’ between the ra’ and the kaf,
Spoil
e On the final line, the Leaks wa-tama ‘an, “and hope,”
of Qi3:12 has also been written over a taping.
IN ALL THESE CASES, what was first written under the tape
cannot be discerned, but likely remains there and could be
seen were the tape carefully removed. It is possible that what
is written underneath matches what was written over the top,
but it is not certain that such is the case. Given the many other
instances of correction in Qur’a4n manuscripts, I have not
ruled out the possibility that some of these tapings are
covering with intent to obscure a variant text or, in the case of
tapings overwritten, to change what was first written.
ro
. Altikulag, Tayyar, ed., Al-Mushaf al-Sharif attributed to ‘Uthman bin ‘Affan
(The copy at the Topkapi Palace Museum) (Istanbul: IRCICA, 2007), 5-13.
2. Ibid., 10-13.
. Déroche, Francois, Qur’ans of the Umayyads: A first overview (Leiden: Brill,
2014), 17.
. Déroche, Francois, La transmission écrite du Coran dans les débuts de I’Islam:
Le codex Parisino-petropolitanus (Leiden: Brill, 2009), 173; Déroche, Francois,
Qur ans of the Umayyads: A first overview (Leiden: Brill, 2014), 34.
. Ibid.
. Altikulag, Tayyar, al-Mushaf al-Sharif Attributed to ‘Uthman bin Affan: The
Copy At al-Mashhad al-Husayni in Cairo (Istanbul: IRCICA, 2009), 131-3.
.. Throughout this book, I transcribe the archigrapheme A LLH as allah. The
use of d instead of a adds an element (the presumption of a long vowel)
that is not, strictly speaking, present in the manuscripts.
. Déroche, Francois, Qur’ans of the Umayyads: A first overview (Leiden: Brill,
2014), 96.
Io.
13.
15.
16.
17.
18.
19.
20.
21.
22.
Corrections in Early Qur’an Manuscripts 91
. Déroche, Frangois, La transmission écrite du Coran dans les débuts de l’islam:
Le codex Parisino-petropolitanus (Leiden: Brill, 2009), 10ff.
Déroche, Francois, Qur’ans of the Umayyads: A first overview (Leiden: Brill,
2014), 75-7 154-5.
. Ibid., 105.
12.
In modern convention around this particular word, even this way of
writing it (ie. with the waw) is transliterated salat, but as per my remarks
on transliteration at the front of this book, I am breaking with standard
shorthand in order to precisely represent the script as it appears on the
page.
“Birmingham Qur’an manuscript dated among the oldest in the world,”
University of Birmingham, posted 22 July 2015, https://www.birmingham.
ac.uk/news/latest/2015/07/quran-manuscript-22-07-15.aspx
. Sadeghi, Behnam and Uwe Bergmann, “The Codex of a Companion of the
Prophet and the Qur’an of the Prophet,” in Arabica 57 (January 2010): 343-
436. See also Sadeghi, Behnam and Mohsen Goudarzi, “San‘a 1 and the
Origins of the Qur’an,” in Der Islam 87 (March 2012): 1-129.
Brubaker, Daniel, “Asking Forgiveness Seventy Times,” (conference paper,
Middle East Studies Association Annual Meeting, San Diego, CA,
November, 2010).
I place omission in quotation marks because the fact that the particular
page to which I refer was corrected but these words were not added
requires us to at least consider whether these words were felt to belong
here at the time and place of both original production and correction of
this manuscript.
Cook, Michael, “The stemma of the regional codices of the Koran,” in
Graeco-Arabica: Festschrift in honor of V. Christides Tipmtikoa Topoo BaotAetov
Kpnottén (Athens: Graeco Arabica, 2004), 93-4. There is more to say about
this, as this particular change reflects something that has been discussed
extensively in the literatures from the time. For the purposes of this book,
however, it is enough to know this.
Thanks to Marijn van Putten for pointing out this explanation. I have
looked at this correction for years — of course my attention is on thou-
sands of pages and not intensely upon this one alone — without realizing
that this was the function of the oval mark here.
Déroche, Francois, Catalogue des manuscrits arabes : deuxiéme partie : manu-
scrits musulmans : tome I, t (Paris: Bibliotheque nationale, 1983), 63-69.
Déroche, Frangois, Qur’ans of the Umayyads: A first overview (Leiden: Brill,
2014), 80.
Ibid., 76.
Its style is very close to that of CBL Is 1615 I/II in Dublin, with hand almost
identical, van Putten has observed. (personal communication)
92
DANIEL ALAN BRUBAKER
23. Stewart, Devin, “Divine Epithets and the Dibacchius: Clausulae and
24.
25.
26.
Qur’anic Rhythm,” in Journal of Quranic Studies 15.2 (2013): 22-64. Stewart
has done good work on rhyme patterns, asking whether current readings
may in some cases not be the original readings. I was riveted when I first
heard him present a paper on this several years ago, and I believe the line
of inquiry holds potential as an item in the toolbox during the coming
years of manuscript research.
Déroche, Frangois, Catalogue des manuscrits arabes : deuxiéme partie : manu-
scrits musulmans : tome I, 1 (Paris: Bibliotheque nationale, 1983), 67.
Brubaker, Daniel, “Frequently Corrected Verses In Early Qur’an Manu-
scripts,” (paper presented at the Annual Meeting of the European Associa-
tion of Biblical Studies, Leuven, Belgium, July 2016).
Brubaker, Daniel, “Corrections involving the word rizq (“provision”) in
early Qur’ans,” (paper presented at the Annual Meeting of the
International Qur’anic Studies Association, San Antonio, TX, November
2016).
27. Rubin, Uri, “Meccan trade and Qur’anic exegesis (Qur’an 2:198),” in
28.
29.
30.
31.
32.
33.
34.
Bulletin of the School of Oriental and African Studies, University of London 53
no. 3 (1990), 421-428.
Déroche, Francois, Catalogue des manuscrits arabes : deuxiéme partie : manu-
scrits musulmans : tome I, 1. (Paris: Bibliotheque nationale, 1983), 109, 120,
131, 138.
Déroche, Francois. The Abbasid Tradition: Qur’ans of the 8th to the roth
Centuries AD (London: Nour Foundation, 1992), 54-55.
Déroche, Frangois, Catalogue des manuscrits arabes : deuxiéme partie : manu-
scrits musulmans : tome I, 1. (Paris: Bibliotheque nationale, 1983), 69.
Ibid., 147.
Alukulag, Tayyar, al-Mushaf al-Sharif Attributed to ‘Uthman bin Affan: The
Copy At al-Mashhad al-Husayni in Cairo, 2 vols. (Istanbul: Organisation of
the Islamic Conference Research Centre for Islamic History, Art, and
Culture (IRCICA), 2011), 124-5.
Rahman, Fazlur, Major Themes of the Qur’an (Chicago: The University of
Chicago Press, 2009), 106ff; Cook, David, Contemporary Muslim Apocalyptic
Literature (Syracuse: Syracuse University Press, 2005), 8-9.
Altikulag, Tayyar, al-Mushaf al-Sharif Attributed to ‘Uthman bin Affan: The
Copy At al Mashhad al-Husayni in Cairo, 2 vols. (Istanbul: Organisation of
the Islamic Conference Research Centre for Islamic History, Art, and
Culture (IRCICA), 2011), 124-5.
3
CONCLUSIONS
he Qur’an has been, and continues to be,
consequential in the affairs of men. In many parts of
the world, it is a source of regional, cultural, and
spiritual pride inextricably intertwined with every part of life.
It is also an object of history related to one of the most
dramatic and enduring movements of political conquest and
colonization in the history of the world. It claims internally
(e.g. Q2:1) to be revelation from God, and was also claimed as
such by Muhammad himself. Furthermore, as a piece of
writing (Arabic kitab) with poetic and linguistic nuance, allu-
sions to events and details of its time as well as to the biblical
scriptures (Hebrew Bible and New Testament) and apocryphal
writings, it contains theological and historical themes inter-
twined in complex ways. For all these reasons and more, it is
an object that has attracted scholarly study from many
different directions.
Leaving aside for the moment devotional considerations
— because these are generally outside the scope of an acad-
94. DANIEL ALAN BRUBAKER
emic inquiry — there are many ways to approach the history
of the Qur’an. For example, there is analysis through the lens
1 and
of secondary literatures, both Arab/Muslim sources
others,” both of which can carry special problems, including
internal or external contradictions;> there is linguistic? and
poetic or chiastic® analysis of the words and word groupings®
of the Qur’an itself, or of the presence of foreign words;’ there
is the Qur’ans self-referentiality,® there is study of the histor-
ical content and clues in the text of the Qur’an, such as places,
people, and references to historical events and topography,”
there is consideration of the theological and legal themes and
motifs of the Qur’an in context of its time and place of deliv-
ery;!° and more.
Then, there is analysis of the material history,!! which
includes physical traces of Qur’an passages, such as in rock
inscriptions or on monuments from the early time periods.
This includes consideration of the political circumstances in
the period following the lifetime of Muhammad.’
Factoring large in the material history, of course, is the
manuscripts, which serve as witnesses to the both their time of
first production and also to the time (if applicable) of
correction.
In the preceding pages, I have shown examples of correc-
tions from Qur’an manuscripts that were produced in the first
several centuries after the death of Muhammad. As stated
earlier, I did not choose the most dramatic examples to
present here, but rather a good group of samples to introduce
the range of the phenomenon. In order to provide readers the
most value, I’ve generally decided not to pick corrections that
I’ve judged to be the result of correcting a mere scribal mistake
from the time of first production; the one exception in this
Corrections in Early Qur’an Manuscripts 95
book is (possibly) Example 8. Among all the corrections I’ve
documented so far in my research, simple scribal error does
account for some of them, and it is important for readers to
understand that this explanation is the first factor I consider
when trying to discern the cause. These manuscripts were
written by human beings, not machines, and so ordinary
human error must always be taken into account.
Qa”
WuaT DOES the existence of these corrections mean? It is an
open-ended question with many possible answers. Here are a
few of my thoughts:
First, although it seems to have been reasonably demon-
strated by now that (with the exception of the lower layer of
the San‘a’ palimpsest) most surviving Qur’an manuscripts
bear the signs of having been produced following a campaign
of standardization basically consistent with that reported to
have been directed by the third caliph, it is also clear that there
existed some differences of perception about the correct words
of the Qur’an text at the times most of these manuscripts were
produced, which were later revisited when these perceptions
changed or standardization became more thorough. It is not
impossible that some of these varying perceptions would have
been tied to certain geographic regions or locales. This
perceived flexibility exceeds the bounds of what is reported in
the gird at literature.
Second, these differences of perception were not confined
to the earliest decades after Muhammad’s death, but there was
some flexibility extending for several centuries after. The flexi-
bility does not appear to have been great. For example, with
96 DANIEL ALAN BRUBAKER
few exceptions like the 7th century San‘a’ and Birmingham
palimpsests, we do not usually see the correction of very large
portions of Qur’an text in the manuscripts. This degree of
apparent flexibility that has limits seems to fit very well with
what is seen elsewhere, such as the inscriptions in the Dome
of the Rock which suggest to Chase Robinson and Stephen
Shoemaker a certain instability in the text of the Qur’an
through the time of its completion in 691/2 AD, during the
reign of the caliph ‘Abd al-Malik,'° and, to the larger point,
the variations requiring later correction in the manuscripts
would be consistent with what Nicolai Sinai has termed the
‘emergent canon model,’ the hypothesis that “the Qur’anic
text, in spite of having achieved a recognizable form by 660,
continued to be reworked and revised until c. 700.”!* Of
course, such a model, i.e. complete closure of the quranic
“canon” around 700, would still fail to account for manuscripts
being produced after this time that still required later correc-
tion, unless of course every one of these were to be attributed
only to orthographic developments, standard qira’at varia-
tions, or scribal error at first production, a scenario that does
not appear to be the case.
Third, partial correction suggests a movement toward a
standard over time, a gradual process rather than a sudden
complete standardization. By partial correction, | mean places
where one aspect of the writing on a page was brought to
conformity with the 1924 Cairo rasm but another part of the
writing remained uncorrected. Of course, this surmise
suggests that the corrector, when noting and revising one
aspect of the writing on the page that he perceived to be
deviant, passed over another that he presumably did not see to
be incorrect.
Corrections in Early Qur’an Manuscripts 97
A dominant traditional view about the Qur’an’s early
transmission and preservation held that orality was the
primary factor, and the ability of even modern children to
memorize the entire Qur’an from an early age is held forth as
evidence that the same was the practice during the time of
Muhammad and the centuries following. Indeed, there is little
reason to doubt that oral transmission played a significant role
in those early years. However, the existence of manuscripts
attest also to a tradition of written transmission, and features
of the manuscripts also suggest the practice of scribal copying
from an exemplar.!> That is, they looked at an existing copy in
order to make a new copy, rather than either writing from
memory or writing from hearing a recitation.!*° So, it is more
likely that orality was part of the picture but that the major
transmission of the book was not purely oral, an environment
that Sadeghi and Bergmann have termed “semi-orality.”!7
A reconstruction of the physical history of the manuscripts
and their relationship to both the oral tradition(s) and to one
another is one goal of this work. There is, in particular, the
hope of grouping manuscripts into families based upon close
analysis and their textual features; this area of research is
called stemmatics, and it highlights the familial relationship
from parents (the exemplar) to children (the copies), grand-
children, cousins, and so forth. It should come as no surprise
that this biological model should make use of methods and
tools employed in similar work in the area of biology, and Alba
Fedeli, for example, has been conducting analysis in this
way.'® The larger idea is one that has long been employed in
biblical textual criticism and is well-developed in that field.
Nor is it a novel concept when it comes to the Qur’an manu-
scripts; classifications of these objects into families was
98 DANIEL ALAN BRUBAKER
proposed by Theodore Néldeke as early as 1860,'? and others
have used distinctive features as a means of grouping manu-
scripts according to relational proximity.”°
Clearly, the above observations bear only upon the trans-
mission of the Qur’an. That is to say, they do not have
anything to say about questions of whether Muhammad
received revelation or whether this revelation was from God;
rather, they speak only to what happened later as the commu-
nity of believers preserved and passed along what he delivered
to them.
The mere existence of corrections in manuscripts is not the
end of the story but a piece of the picture that must be taken
into account when assessing what was being transmitted, in
this case the words of what came to be understood by believers
in Muhammad's apostleship to be a set of revelations from
God. A manuscript is a physical record of a text; it isa medium
of transmission and of preservation. We have many ways of
transmitting and preserving information in recent years and
today: print, photography, magnetic recordings such as
cassette and VHS tapes, CDs and DVDs, digital archives, and
of course (as in the 7th century) the handwritten document. In
each case there is the possibility of noise or distortion caused
by either human error or the limitations of the medium itself,
but not every variation between records is necessarily the
result of human error or the limitations of the medium. The
work of a manuscript researcher is work in the real world of
objects, using judgment to discern what is noise and what is
meaningful information. I have drawn only a few conclusions
here but expect that in the end, the greatest value of this book
will have been the opportunity for reflection that the
photographs and descriptions has provided you.
Corrections in Early Qur’an Manuscripts 99
Certainly, there is much more to be said, and a great
amount of material remains for further scholarly research. I
will continue as I am able and hope others will also.
1. These include early literatures that fall into various categories: tafsir (com-
mentary), tarikh (history), sira (biography - i.e. of Muhammad), rijal (liter-
ally “men,” it is literature about the lives, lineages, and reputations for
truthfulness and character of the people who were involved in transmitting
traditions), hadith (accounts of “what happened,” organized topically and
in discrete bits of information as reportedly passed from person to person
until being collected and written down by the hadith collector, maghazi
(histories of raids and conquests), fiqh (legal texts rooted in the teachings of
Muhammad and the Qur’an), to name a few. Needless to say, the earlier
ones tend to carry a special weight with scholars even if they are not in
every case the most popular devotionally. Also, there are some works that,
for various reasons, are considered to be more authoritative than others.
Even the most authoritative works are not without their problems, and this
is partly because all of these works tend to be separated from the events
they describe by more than a century.
2. For example: Hoyland, Robert G., Seeing Islam as others saw it: a survey and
evaluation of Christian, Jewish, and Zoroastrian writings on early Islam (Prince-
ton: The Darwin Press, 1997).
3. Rippin, Andrew, “Al-Zuhri, Naskh al-Qur’an and the problem of early tafsir
texts,” in Bulletin of the School of Oriental and African Studies, University of
London 47 no. 1 (1984), 22-43; Donner, Fred McGraw, The Early Islamic
Conquests, (Princeton: The Princeton University Press, 1981); Motzki, Harald,
“Whither Hadith Studies?” in Analysing Muslim Traditions: Studies in legal,
exegetical, and maghazi hadith (Leiden: Brill, 2010), 47-124; Crone, Patricia,
Meccan Trade and the Rise of Islam (Piscataway: Gorgias Press, 2004); Noth,
Albrecht, The early Arabic historical tradition: A source-critical study (Prince-
ton: The Darwin Press, 1994); Neuwirth, Angelika, “Qur’an and History —
a Disputed Relationship: Some reflections on Qur’anic History and
History in the Qur’an,” in Journal of Quranic Studies 5 no. 1 (2003), 1-18;
Crone, Patricia, “How did the quranic pagans make a living?” in Bulletin of
the School of Oriental and African Studies, University of London 68 no. 3 (2005),
387-399.
4. Luxenberg, Christoph, The Syro-Aramaic Reading of the Koran: A contribution
to the decoding of the language of the Koran (Berlin: Verlag Hans Schiler,
100
10.
Il.
DANIEL ALAN BRUBAKER
2007); Durie, Mark, The Qur’an and its biblical reflexes (Lanham: Lexington
Books, 2018).
. Cuypers, Michel, The Banquet: A reading of the fifth sura of the Qur’an
(Miami: Convivium Press, 2009); Cuypers, Michel, A Quranic Apocalypse: A
reading of the thirty-three last surahs of the Qur’an (Atlanta: Lockwood Press,
2018); Stewart, Devin, “Divine Epithets and the Dibacchius: Clausulae and
Qur’anic Rhythm,” in Journal of Qur’anic Studies, 15.2 (2013), 22-64; Rippin,
Andrew, “The poetics of Qur’anic punning,” in Bulletin of the School of
Oriental and African Studies, University of London 57 no. 1 in Honour of J. E.
Wansbrough (1994), 193-207.
. Bannister, Andrew G., An Oral-Formulaic Study of the Qur’an (Lanham:
Lexington Books, 2014); Witztum, Joseph, “Variant Traditions, Relative
Chronology, and the Study of Intra-Quranic Parallels,” in Islamic Cultures,
Islamic Contexts: Essays in honor of Professor Patricia Crone, ed. Behnam
Sadeghi, Asad Q. Ahmed, Adam Silverstein, and Robert Hoyland (Leiden:
Brill, 2015); Durie, Mark, “Phono-semantic matching in Qur’anic Arabic,”
(unpublished paper, Arthur Jeffery Centre for Islamic Studies, Melbourne
School of Theology).
. Jeffery, Arthur, The Foreign Vocabulary of the Qur’an (Leiden: Brill, 2007).
. Madigan, Daniel A., The Qur’dn’s self-image: Writing and authority in Islam's
Scripture (Princeton: Princeton University Press, 2001).
. Zellentin, Holger Michael, The Qur’an’s Legal Culture: The Didascalia Apos-
tolorum as a Point of Departure (Tiibingen: Mohr Siebeck, 2013).
Dost, Suleyman, “An Arabian Qur’an: Towards a theory of peninsular
origins,” (PhD diss., University of Chicago, June 2017); “Geiger, Abraham,
Was hat Mohammed aus dem Judenthume aufgenommen? (Berlin: Parerga,
2005); Reynolds, Gabriel Said, The Qur’an and Its Biblical Subtext (Abing-
don: Routledge, 2010); Reynolds, Gabriel Said, ed., The Quran in Its Histor-
ical Context (Abingdon: Routledge, 2008); Reynolds, Gabriel Said, ed., New
Perspectives on the Qur’an: The Quran in its historical context 2 (Abingdon:
Routledge, 2011); Zellentin, Holger Michael, The Qur’an’s Legal Culture: The
Didascalia Apostolorum as a Point of Departure (Tiibingen: Mohr Siebeck,
2013).
Small, Keith E., Textual Criticism and Qur’dn Manuscripts (Lanham:
Lexington Books, 201); Fedeli, Alba, “Early Qur’anic manuscripts, their
text, and the Alphonse Mingana papers held in the Department of Special
Collections of the University of Birmingham,” (PhD diss., University of
Birmingham, 2014); Powers, David, Muhammad is not the father of any of your
men (Philadelphia: University of Pennsylvania Press, 2009); Puin, Elisabeth,
“Ein frither Koranpalimpsest aus San‘a’? (DAM o1-27.1),” in Schlaglichter:
Die beiden ersten islamischen Jahrhunderte, ed. Gro8&, Markus and Karl-Heinz
Ohlig, (Berlin: Verlag Hans Schiler, 2008); Dutton, Yasin, “Some Notes on
12.
13.
14.
15.
16.
17.
18.
19.
Corrections in Early Qur’an Manuscripts IOI
»
the British Library’s ‘Oldest Qur’an Manuscript’ (Or. 2165),
Quranic Studies 6 no. 1 (2004), 43-71; Sadeghi, Behnam and Uwe Bergmann,
in Journal of
“The Codex of a Companion of the Prophet and the Qur’an of the
Prophet,” in Arabica 57 (2010), 343-436; Rezvan, E., “New folios from ‘‘Uth-
manic Qur’an’ I. (Library of Administration for Muslim Affairs of the
Republic of Uzbekistan),” in Manuscripta Orientalia 10 no. 1 (2004). These
are just a sampling from a much wider pool of work, including works cited
earlier in this book.
Kohlberg, Etan, and Mohammad Ali Amir-Moezzi, eds., Revelation and Falsifi-
cation: The Kitab al-qira’at of Akmad b. Muhammad al-Sayyani (Critical Edition)
(Leiden: Brill, 2009; Modarressi, Hossein, “Early Debates on the Integrity of
the Qur’an: A Brief Survey,” in Studia Islamica 77 (1993), 5-39. There were
early debates, for example, in which it was alleged that the commonly
accepted text of the Qur’an had been corrupted. The book mentioned here
is a critical edition of one such work from the gth century AD.
Sinai, Nicolai, “When did the consonantal skeleton of the Qur’an reach
closure?” in Bulletin of the School of Oriental and African Studies 77 (2014),
273-292.
Ibid. 6.
Clues that a manuscript has been copied by looking at another manuscript
include mistakes such as haplography (omitting a word or phrase) or dittog-
raphy (writing the same word or phrase twice) due to parablepsis (looking
aside when copying, for example, to fill the ink in a nib). There are
numerous instances of correction in early Qur’4n manuscripts that rectify
this kind of mistake.
This latter practice, writing from recitation, could be discerned when, for
example, letters that sound the same but look different when written are
interchanged. Such a mistake would not be made if the scribe was copying
from an earlier manuscript. This sort of mistake is not common in Qur’an
manuscripts; in fact, no example of it comes to mind.
Sadeghi, Behnam and Uwe Bergmann, “The Codex of a Companion of the
Prophet and the Qur’an of the Prophet,” in Arabica 57 (2010), 345.
Fedeli, Alba, and Andrew Edmondson, “Early Qur’anic Manuscripts and
their Networks: a Phylogenetic Analysis Project,” (pre-circulated paper for
Conference “Qur’anic Manuscript Studies: State of the Field,” Budapest,
May 2017, after the research project Early Qur’anic Manuscripts and their
Relationship as Studied Through Phylogenetic Software at the Central
European University, Budapest).
Cook, Michael, “The stemma of the regional codices of the Koran,” in
Graeco-Arabica Festschrift in Honour of V. Christides Tiypntixoo Toptoo BactAgtov
Xpnotidn, Volumes IX:X. ed. George Livadas. (Athens: Graeco Arabica,
2004), 89-104.
102 DANIEL ALAN BRUBAKER
20. George, Alain, “Coloured Dots and the Question of Regional Origins in
Early Qur’ans (Part [),” in Journal of Qur’anic Studies 17.1 (2017), 1-44; van
Putten, Marijn, “The Grace of God’ as evidence for a written Uthmanic
Archetype: The importance of shared orthographic idiosyncrasies,” in
Bulletin of the School of Oriental and African Studies (forthcoming).
INDEX OF QUR’ AN VERSES REFERENCED
2:137 (62-3); 2:191-3 (85, 87-8); 3:171 (66-7); 4:6 (76-7); 4:33 (75-6);
4:149 (58-9); 4:167 (70-2); 5:93 (52-4); 6:40 (79-80); 6:86 (67); 6:91-
97 (47-51); 7:34 (79-80); 8:3 (83-4); 9:72 (28-30); 9:78 (34, 38-9, 43);
9:80 (56); 9:93 (34, 38, 43); 13:11-12 (89-90); 15:85 (78-80); 22:40 (34,
37, 41); 23:86 (55-7); 23:87 (56-7); 24:33 (73-4); 24:51 (34, 37, 42); 30:9
(44-6); 33:9 (78-79); 33:18 (34, 36, 39); 33:24 (34, 36, 40); 33:73 (34,
36-7, 40); 34:27 (81-2), 34:35 (68-69); 35:11 (34, 37, 42); 41:21 (34, 37,
41); 42:5 (60-1); 42:21 (31-3); 66:8 (64-5)
FURTHER READING
Below is a partial list of recent books specifically dealing with
Qur’an manuscripts. Some may be challenging for a non-
specialist. My mention is not an endorsement of every posi-
tion, theory, or conclusion of the authors, but all are serious
scholars engaging substantively with the subject.
Baker, Colin F. Qur’an manuscripts: calligraphy, illumination,
design. London: The British Library, 2007.
Blair, Sheila S. Islamic calligraphy. Edinburgh: Edinburgh
University Press, 2008.
Cellard, Eléonore. Codex Amrensis I (French and Arabic).
Leiden: Brill, 2018.
Déroche, Francois. Qur’ans of the Umayyads. Leiden: Brill,
2014.
106 Further Reading
George, Alain. The Rise of Islamic Calligraphy. London: SAQI,
2010.
Hilali, Asma. The Sanaa Palimpsest: The Transmission of the
Qur’an in the First Centuries AH. Oxford: Oxford University
Press, 2017.
Powers, David. Muhammad is not the father of any of your men.
Philadelphia: University of Pennsylvania Press, 2009.
Sinai, Nicolai. The Qur’an: A Historical-Critical Introduction.
Edinburgh: Edinburgh University Press, 2017.
Small, Keith. Textual Criticism and Qur’an Manuscripts.
Lanham: Lexington Books, 2011.
I further recommend the chapters from the following authors
that are contained, among other places, in the German Inarah
volumes edited by Karl Heinz-Ohlig and Markus Grof::
Alba Fedeli
Thomas Milo
Elisabeth Puin
Gerd-R Puin
Keith Small
Others who have published important journal articles, but not
Further Reading 107
yet books, on Qur’an manuscripts include notably Yassin
Dutton, Mohsen Goudarzi, Efim Rezvan, Behnam Sadeghi,
Ahmad Al-Jallad, Michael Marx, and Marijn van Putten. To
this list could be added most of the authors of books above.
Finally, ] mention one additional recently-published book that
does not relate directly to the manuscripts, but which engages
in fine linguistic and thematic analysis that may have some
bearing on some of the things we see going on in them:
Durie, Mark. The Quran and Its Biblical Reflexes: Investigations
Into the Genesis of a Religion. London: Lexington Books, 2018.
GLOSSARY
A.H. (or AH) — the abbreviation for Anno Hegirae, “Year of the
Hijrah,” a designation of a date given according to the Islamic
calendar, which counts lunar years from the time of Muham-
mad’s emigration from Mecca to Medina in 622 AD. Centuries
are often given in AD/AH format, for example “7th/Ist centu-
ry,’ meaning the 7th century AD, which is also the Ist
century AH.
archigrapheme — a mark which can represent different
graphemes (letters). In this context it refers to unpointed
Arabic letters, which, for lack of diacritics, were often
ambiguous.
aya — averse of the Qur’an
bifolio — a sheet folded in the middle so as to form two folios
in a bound book. Several bifolios are usually stacked and sewn
together to form a quire.
IIO Glossary
codex — a book (i.e. multiple pages bound at one edge). The
Arabic word for codex/book is mushaf (pronounced “moos-
hoff,” NOT “mush-off’)
colophon — a statement, usually included at the end of the
book, that contains details about its production. In a Qur’an
manuscript, acolophon might include the name of the callig-
rapher, the date the project was completed, and perhaps the
name of the patron who commissioned it. Unfortunately, the
earliest Qur’an manuscripts do not include colophons.
consonantal skeletal text — the Arabic rasm, that is, the core
structure of written Arabic without any dots or other marks to
disambiguate letters
diacritics — the graphic marks (usually dots today) that
distinguish an otherwise ambiguous consonant. In Arabic, for
example, three dots above a consonantal tooth indicate the
letter tha’, two dots above indicate a ta’, one dot above indi-
cates nun, one dot below indicates ba’, and two dots below
indicates ya’. There are many other examples.
folio — a page in a codex. A folio has a recto (front) side and a
verso (back) side.
grapheme — the smallest unit of a writing system in a
language. This term is relevant to understanding the word
archigrapheme above.
hadith — a report that has been passed from person to person
over time before being written down. Hadith typically tell of
Glossary III
things Muhammad said or did, approved or disapproved, or
similar things that his companions did or said. Separate
hadith reports have been gathered into authoritative
collections.
manuscript — a handwritten document
mushaf (plural: masahif) — the Arabic word for a book
orthography — from the Greek meaning “right writing,” this
refers to the rules for correctly writing a word, particularly its
spelling
parchment — animal skin prepared to receive writing. Parch-
ment is sometimes also called vellum; they are not exact
synonyms, but the terms are often used interchangeably.
gibla — the direction of Islamic prayer, today toward Mecca
quire — a section of a book consisting, usually, of several bifo-
lios stacked and sewn together in the middle. In traditional
bookbinding, and even in quality bindings today, a number of
quires are first produced and then sewn or glued together to
make a complete book.
rasm — an Arabic word describing the bare consonantal
Arabic text. Full written Arabic today has marks representing
consonants, marks representing long vowels, and sometimes
marks to represent short vowels. The rasm refers to the first
two items, but not to the last one.
112 Glossary
OD,
recto — the front side of a folio in a book, abbreviated “r’;
when referring to manuscripts in this book, for example, 26r
means “26 recto,” or the front side of the 26th folio. The other
side is called “verso.”
surah — a chapter of the Qur’an
script grammar — a term coined in 2002 by Thomas Milo to
refer to slight variations in the consonantal skeletal text that
permits disambiguation of some consonants even in the
absence of dots
verso — the back side of a page in a book, abbreviated “v’;
when referring to manuscripts in this book, for example, 26v
means “26 verso,” or the back side of the 26th folio. The other
side is called “recto.”
ABOUT THE AUTHOR
Dr. Brubaker examining folios of a 7th century Qur’an in the Sabah
Collection of the Dar Museum, Kuwait, 2015.
DANIEL BRUBAKER became fascinated by corrections in
Qur’an manuscripts during his Ph.D. work at Rice University,
so fascinated that he chose to make these his prime focus. His
dissertation, titled “Intentional Changes in Qur’an Manu-
scripts” (2014), is the first extensive survey of physical correc-
tions in early written Qur’ans. Its contents and additional
material are forthcoming. This is Brubaker’s first book.
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Open Library
American Libraries
TV News
Understanding 9/11